Petition for Writ of Certiorari — Okello T. Chatrie, Petitioner v. United States
Supreme Court briefJul 28, 2025
Ask Donna
What actually matters in this document.
Text
No. ____
IN THE
Supreme Court of the United States
_________
OKELLO CHATRIE,
Petitioner,
v.
UNITED STATES OF AMERICA,
Respondent.
________
On Petition for a Writ of Certiorari
to the United States Court of Appeals
for the Fourth Circuit
________
PETITION FOR A WRIT OF CERTIORARI
________
GEREMY C. KAMENS
Federal Public Defender
LAURA J. KOENIG
Assistant Federal Public
Defender
PATRICK BRYANT
Assistant Federal Public
Defender
OFFICE OF THE FEDERAL
PUBLIC DEFENDER,
EASTERN DISTRICT OF
VIRGINIA
1650 King Street,
Suite 500
Alexandria, VA 22314
(703) 600-0800
ADAM G. UNIKOWSKY
Counsel of Record
LAUREL A. RAYMOND
JENNER & BLOCK LLP
1099 New York Ave., NW
Suite 900
Washington, DC 20001
(202) 639-6000
AUnikowsky@jenner.com
i
QUESTION PRESENTED
This case concerns the constitutionality of geofence
warrants. For cell phone users to use certain services,
their cell phones must continuously transmit their exact
locations to their service providers. A geofence warrant
allows law enforcement to obtain, from the service
provider, the identities of users who were in the vicinity
of a particular location at a particular time.
In this case, law enforcement obtained, and served on
Google, a geofence warrant seeking anonymized location
data for every device within 150 meters of the location
of a bank robbery within one hour of the robbery. After
Google returned an initial list, law enforcement sought—
without seeking an additional warrant—information
about the movements of certain devices for a longer,
two-hour period, and Google complied with that request
as well. Then—again without seeking an additional
warrant—law enforcement requested de-anonymized
subscriber information for three devices. One of those
devices belonged to petitioner Okello Chatrie. Based on
the evidence derived from the geofence warrant,
petitioner was convicted of armed robbery.
The questions presented are:
1. Whether the execution of the geofence warrant
violated the Fourth Amendment.
2. Whether the exclusionary rule should apply to
the evidence derived from the geofence warrant.
ii
RELATED PROCEEDINGS
This case arises from and is related to the following
proceedings in the United States District Court for the
Eastern District of Virginia and the United States Court
of Appeals for the Fourth Circuit:
•
United States v. Chatrie, 590 F. Supp. 3d 901
(E.D. Va. 2022).
•
United States v. Chatrie, 107 F. 4th 319 (4th Cir.
2024) (panel opinion)
•
United States v. Chatrie, 2024 WL 4648102 (4th
Cir. 2025) (granting rehearing en banc)
•
United States v. Chatrie, 136 F. 4th 100 (4th Cir.
2025) (en banc opinion)
iii
TABLE OF CONTENTS
QUESTION PRESENTED .............................................. i
RELATED PROCEEDINGS.......................................... ii
TABLE OF APPENDICES ............................................ v
TABLE OF AUTHORITIES ......................................... vi
PETITION
FOR
A
WRIT
OF
CERTIORARI ....................................................... 1
OPINIONS BELOW ......................................................... 1
JURISDICTION ................................................................ 1
RELEVANT
CONSTITUTIONAL
PROVISION ........................................................... 1
INTRODUCTION ............................................................. 1
STATEMENT OF THE CASE ....................................... 7
A.
Geofence Warrants..................................... 7
B.
Factual Background ................................. 11
C.
District Court Proceedings ..................... 12
D.
Fourth Circuit Proceedings. ................... 13
REASONS
FOR
GRANTING
THE
PETITION ............................................................ 19
I.
LOWER COURTS ARE DIVIDED
ON THE CONSTITUTIONALITY OF
GEOFENCE WARRANTS. .............................. 19
II.
THIS
CASE
WARRANTS
SUPREME COURT REVIEW......................... 23
A.
The Issue is Important. ........................... 23
B.
III.
IV.
iv
This Case Is an Appropriate
Vehicle........................................................ 27
PETITIONER’S
FOURTH
AMENDMENT
RIGHT
WAS
VIOLATED. ......................................................... 29
A.
The Government Conducted a
Search. ........................................................ 29
B.
The Warrant Violated the
Fourth Amendment. ................................ 32
THE GOOD-FAITH EXCEPTION
DOES NOT APPLY. ........................................... 34
CONCLUSION ................................................................ 37
v
TABLE OF APPENDICES
Appendix A
United States v. Chatrie, 136 F.4th 100 (4th
Cir. 2025) (on rehearing en banc) ............................. 1a
Appendix B
United States v. Chatrie, No. 22-4489, 2024
WL 4648102 (4th Cir. Nov. 1, 2024) (order
granting rehearing en banc) ................................. 143a
Appendix C
United States v. Chatrie, 107 F.4th 319 (4th
Cir. 2024) .................................................................. 145a
Appendix D
United States v. Chatrie, 590 F.Supp.3d 901
(E.D. Va. 2022) ........................................................ 264a
vi
TABLE OF AUTHORITIES
CASES
Carpenter v. United States, 585 U.S. 296
(2018) .............................................. 2, 3, 26, 28, 29, 30
Davis v. United States, 564 U.S. 229 (2011) ........ 36, 37
Groh v. Ramirez, 540 U.S. 551 (2004) ...................34-35
Jones v. State, 913 S.E.2d 700 (Ga. 2025) .................. 21
People v. Seymour, 536 P.3d 1260 (Colo.
2023) .................................................................... 22, 23
Riley v. California, 573 U.S. 373 (2014) .................... 26
Smith v. Maryland, 442 U.S. 735 (1979).................... 15
United States v. Leon, 468 U.S. 897 (1984) ......... 13, 36
United States. v. Miller, 425 U.S. 435 (1976) ............ 15
United States v. Rhine, 652 F. Supp. 3d 38
(D.D.C. 2023) ........................................................... 23
United States v. Smith, 110 F.4th 817 (5th
Cir. 2024), petition for cert. filed, No. 247237 (U.S. May 19, 2025) .......... 19, 20, 28, 30, 31, 32
Wells v. State, No. PD-0669-23, -- S.W.3d --,
2025 WL 980996 (Tex. Crim. App. Apr.
2, 2025) ................................................................ 20, 21
Wyoming v. Houghton, 526 U.S. 295 (1999).............. 36
CONSTITUTIONAL PROVISIONS AND STATUTES
U.S. Const. amend. IV ................................................. 32
18 U.S.C. § 924(c)(1)(A)................................................ 12
18 U.S.C. § 2113(a) ........................................................ 12
vii
18 U.S.C. § 2113(d)........................................................ 12
18 U.S.C. § 2113(e) ........................................................ 12
28 U.S.C. § 1254(1) .......................................................... 1
OTHER AUTHORITIES
Haley Amster & Brett Diehl, Note, Against
Geofences, 74 Stan. L. Rev. 385 (2022) ................ 24
Andrew Couts, Security News This Week:
Geofence
Warrants
Ruled
Unconstitutional—but That’s Not the
End of It, WIRED (Aug. 17, 2024),
https://www.wired.com/story/geofencewarrants-ruled-unconstitutional-tmobil
e-fine-deepfake-nudes-lawsuit/ ............................. 24
Sidney Fussell, An Explosion in Geofence
Warrants Threatens Privacy Across the
U.S.,
Wired
(Aug.
27,
2021),
https://www.wired.com/story/geofencewarrants-google/ ................................................24-25
Orin S. Kerr, Data Scanning and the Fourth
Amendment (Stan. Pub. L., Working
Paper, 2025), https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=5175686.................. 24
Orin Kerr, The Fifth Circuit Shuts Down
Geofence Warrants—and Maybe A Lot
More, Lawfare (Aug. 14, 2024),
https://www.lawfaremedia.org/article/th
e-fifth-circuit-shuts-down-geofence-war
rants-and-maybe-a-lot-more ................................. 24
viii
Orin S. Kerr, The Fourth Circuit’s
Geofencing Case Ends Not With a Bang
But A Whimper, reason (May 2, 2025),
https://reason.com/volokh/2025/05/02/th
e-fourth-circuits-geofencing-case-endsnot-with-a-bang-but-a-whimper/ .......................... 24
A. Reed McLeod, Note, Geofence Warrants:
Geolocating the Fourth Amendment, 30
Wm. & Mary Bill Rts. J. 531, 532 (2021) .............. 24
Note, Geofence Warrants and the Fourth
Amendment, 134 Harv. L. Rev. 2508
(2021) ........................................................................ 24
Jackie O’Niel, Much Ado About Geofence
Warrants, Harv. L. Rev. Blog. (Feb. 18,
2025),
https://harvardlawreview.org/
blog/2025/02/much-ado-about-geofencewarrants/ .................................................................. 24
Brian L. Owsley, The Best Offense Is A
Good Defense: Fourth Amendment
Implications of Geofence Warrants, 50
Hofstra L. Rev. 829 (2022) .................................... 24
Shira Ovide, Police Love Google’s
Surveillance Data: Here’s How to
Protect Yourself, Wash. Post (Oct. 24,
2023), https://www.washingtonpost.com/
technology/2023/10/24/google-privacy-p
olice-geofence/ ......................................................... 24
ix
Stipulation Order to Partially Unseal
Records in Sealed Case, In re Google,
Case No. 25-mj-70146 (N.D. Cal. Feb. 7,
2025), ECF No. 1, Exhibit A (motion to
quash geofence warrant filed on
February 7, 2024) ...............................................10-11
Jennifer
Valentino-deVries,
Tracking
Phones, Google is a Dragnet for the
Police, N.Y. Times (Apr. 13, 2019),
https://www.nytimes.com/interactive/20
19/04/13/us/google-location-tracking-po
lice.html .................................................................... 25
1
PETITION FOR A WRIT OF CERTIORARI
Petitioner Okello Chatrie respectfully petitions for a
writ of certiorari to review the judgment of the United
States Court of Appeals for the Fourth Circuit.
OPINIONS BELOW
The Fourth Circuit’s en banc opinion is reported at
136 F.4th 100 (4th Cir. 2025). Pet. App. 1a. The Fourth
Circuit’s initial opinion is reported at 107 F.4th 319 (4th
Cir. 2024). Pet. App. 145a. The district court’s opinion
denying suppression is reported at 590 F. Supp. 3d 901
(E.D. Va. 2022). Pet. App. 264a.
JURISDICTION
The Fourth Circuit issued its decision on April 30,
2025. This Court has jurisdiction under 28 U.S.C.
§ 1254(1).
RELEVANT CONSTITUTIONAL PROVISION
The Fourth Amendment provides: “The right of the
people to be secure in their persons, houses, papers, and
effects, against unreasonable searches and seizures,
shall not be violated, and no Warrants shall issue, but
upon probable cause, supported by Oath or affirmation,
and particularly describing the place to be searched, and
the persons or things to be seized.”
INTRODUCTION
This case presents important constitutional
questions concerning the controversial law enforcement
tool known as the “geofence warrant.”
Many cell phone owners use services that
continuously transmit their exact location to their
2
service providers.
Service providers use this
information to track users’ location over time. Based on
this information, service providers can identify all users
who were in a particular geographic area at a given time.
That capability has given rise to the geofence
warrant. Geofence warrants seek information regarding
devices that were within a particular geographic area
during a particular time period. The warrant draws a
virtual “fence” around a particular geographic area
(hence “geofence”) and seeks information about devices
within that “fence” during the relevant time period.
This case involves a geofence warrant served on
Google. Every two minutes, Google’s Location History
service records the user’s location. The geofence
warrant in this case directed Google to scan through the
private user-controlled accounts of over 500 million
Location History users to identify all devices that were,
within one hour of a bank robbery, within 150 meters
from the scene of the crime. After Google complied with
that request, law enforcement sought and received
additional location information for certain devices whose
movements law enforcement deemed suspicious.
Finally, without obtaining an additional search warrant,
law enforcement requested and received names
associated with three devices—which included
petitioner Okello Chatrie. Petitioner was charged with,
and convicted of, bank robbery.
Geofence warrants are a powerful law enforcement
tool. At the same time, they raise significant Fourth
Amendment concerns. This Court has been attuned to
the privacy risks of new law enforcement techniques
involving cell phones: In Carpenter v. United States, 585
3
U.S. 296 (2018), this Court held that cell phone users
retain a reasonable expectation of privacy in the record
of their physical movements as captured by cell-site
location information (CSLI), through which cell
providers can triangulate and track cell phone users’
location. Id. at 309-10. The privacy concerns arising
from geofence warrants are even greater than the
privacy concerns at issue in Carpenter. In Carpenter,
law enforcement officials sought information about the
movements of a single individual suspected of a crime
based on the movements of his cell phone. By contrast,
using a geofence warrant, law enforcement may request
information regarding all people who were at a sensitive
location—an abortion clinic, a protest, a political party’s
convention—at a particular time.
It would be an understatement to say that lowercourt judges have disagreed on the constitutional
questions arising from geofence warrants. The Fourth
Circuit panel in this case held that obtaining location
information from Google was not a search at all. On
rehearing en banc, the Fourth Circuit fractured, issuing
a one-sentence per curiam affirmance accompanied by
nine separate opinions.
The Fifth Circuit, meanwhile, has held both that
requiring Google to search through over 500 million user
accounts for responsive location information is a search,
and that geofence warrants are unconstitutional general
warrants, effectively banning geofence warrants in
federal courts within the Fifth Circuit. The Texas Court
of Criminal Appeals has rejected the Fifth Circuit’s
view—leading to a split between state courts and federal
4
courts in the same jurisdiction—as has the Georgia
Supreme Court.
The uncertainty over geofence warrants is
intolerable. When magistrate judges receive requests
for geofence warrants, they need to know what rules
apply. The same is true for tech companies that wish to
cooperate with law enforcement while also protecting
their users’ privacy and complying with the
Constitution. Although Google announced plans to
change its internal policies such that it no longer stores
location information on its central servers, the state of
those plans and the status of its existing data are both
unclear, and geofence warrants have been served on
other companies as well, including Apple, Lyft,
Snapchat, and Uber. This issue is not going away. This
Court’s review is needed.
On the merits, this Court should hold that the
geofence warrant violated petitioner’s Fourth
Amendment rights. To begin, as Judge Wynn and Judge
Berner’s opinions below explain in detail, Carpenter
dictates the conclusion that the government conducted a
search. The Carpenter Court held that the collection of
CSLI was a search because cell phone users have a
reasonable expectation of privacy in CSLI, and every
aspect of the Court’s reasoning in Carpenter applies
with equal or greater force here. Location History data
is more comprehensive than the CSLI at issue in
Carpenter; Location History allows for surveillance of
everyone in a particular “fence,” rather than just one
person; Location History, like CSLI, allows access to
intimate information about a person’s life; and, as with
5
CSLI, obtaining Location History
inexpensive for law enforcement.
is
easy
and
In Carpenter, Justice Gorsuch stated that the Fourth
Amendment analysis should turn not on whether a
person has a reasonable expectation of privacy, but
instead on whether he has a property interest in the item
searched. Applying that approach here leads to the
same result. Unlike CSLI, Location History is not a
business record. Instead, Location History belongs to
the user—the user can delete it or direct Google to stop
collecting it. In this case, by collecting information that
belonged to petitioner, law enforcement conducted a
search.
Although the government obtained a geofence
warrant, that warrant was inadequate to justify the
search. As the Fifth Circuit has held, geofence warrants
do not satisfy the Fourth Amendment’s particularity
requirement. Even assuming, contrary to the Fifth
Circuit’s view, that geofence warrants are sometimes
constitutional, the execution of the warrant violated the
Fourth Amendment here. The government should have
obtained a warrant after Google returned its list.
Obtaining a warrant at the first step was insufficient
under the Fourth Amendment to justify the
government’s winnowing process and ultimate
unmasking of petitioner’s identity without further court
review.
The Court should therefore grant certiorari on the
first question presented and hold that the execution of
the geofence warrant violated the Fourth Amendment.
6
The Court should also grant review on the second
question presented: whether the exclusionary rule
applies.
In the Fourth Circuit, every judge other than Judge
Gregory concluded either that there was no Fourth
Amendment violation or that the good-faith exception to
the exclusionary rule applied. Petitioner agrees with
Judge Gregory’s assessment that the good-faith
exception does not apply because the officers’ conduct
was not objectively reasonable. Further, the good-faith
exception should not apply where, as here, the warrant
at issue is a general warrant that authorizes the search
of millions of accounts without probable cause.
Additionally, there is a more fundamental reason to
reject the good-faith exception in this case. If the Court
applies the good-faith exception, it will be impossible for
this Court to issue an opinion deciding the
constitutionality of geofence warrants (or any other
novel surveillance tool) that will benefit the litigant. In
every civil case, the officer will be protected by qualified
immunity. In every criminal case, the defendant’s
conviction will stand regardless of the warrant’s
constitutionality. The result will be that confusion and
lower-court splits will persist forever.
This Court has left open the possibility that it would
not apply the exclusionary rule in cases where the goodfaith exception would stunt the development of the law.
That is this case. The Court should hold that the
wrongfully-obtained evidence cannot be used against
petitioner and reverse petitioner’s conviction.
7
STATEMENT OF THE CASE
A. Geofence Warrants
The geofence warrant in this case was directed at
Google and relied on a Google feature called “Location
History.” Pet. App. 270a. Location History draws from
GPS information, Bluetooth beacons, CSLI, IP address
information, and nearby Wi-Fi networks to log a device’s
location, making a record, on average, every two
minutes. Pet. App. 271a. It is powerful enough to
determine if a person is on the “second [or first] floor of
[a] mall,” and can determine a person’s location to within
as little as three meters. Pet. App. 272a, 274a.
Location History is not automatically enabled.
However, Google offers multiple methods to turn
Location History on and typically prompts users to do so
multiple times across multiple apps, starting from when
a user first sets up her Google account. Pet. App. 273a.
If a user does not enable Location History immediately
upon account setup, Google will prompt her to do so
when she sets up an app with “Location Historypowered features,” like Google Maps, Google Photos,
and Google Assistant, telling her it is necessary to “[g]et
the most from” the app. Pet. App. 279a (brackets
omitted). And once a user enables Location History
through one app, it is enabled across all of her Google
devices, and it will continue to be active even if she
deletes the original app. Pet. App. 273a. Google collects
and stores Location History at all times, regardless of
whether the user is actively using her phone.
After opting in, a user can “pause” the collection of
her data, though if she does so, a pop-up screen will warn
her that pausing Location History will “limit[]
8
functionality of some Google products over time.” Pet.
App. 282a (bracket in original). Pausing Location
History does not delete previously-collected data—it
merely pauses collection of new data. Pet. App. 283a. If
a user wanted to delete her location history in 2018—
when petitioner first (seemingly inadvertently) enabled
his Location History—there was only one way to do so:
by visiting myactivity.google.com. Pet. App. 281a, 283a,
333a. One Google employee remarked that the user
interface “*feels* like it is designed to make things
possible, yet difficult enough that people won’t
figure…out” how to turn Location History off—a
sentiment the district court endorsed. Pet. App. 283a284a.
Because Google produces both Android phones and
apps like Google Maps that function across devices,
Google is able to collect detailed location data, including
Location History, on “numerous tens of millions” of
users. Pet. App. 270a; see also Pet. App. 272a n.8. At
the time the warrant in this case was executed, Google
stored this extensive data in a repository known as the
“Sensorvault,” where each datapoint is associated with
a unique device ID. Pet. App. 272a. Google uses this
data for advertising and to power models and support
features of Google apps, such as maps. Pet. App. 272a.
While the default storage method is anonymized, Google
can also de-anonymize the information to reconstruct the
movements of a particular user. Even without explicit
de-anonymization, the granularity of the data is such
that, in many cases, a few data points, used in
conjunction with other publicly available information,
can be used to unmask a user’s likely identity. See Pet.
App. 305a.
9
Geofence warrants rely on this vast trove of data.
When law enforcement serves a geofence warrant on
Google, it identifies a geographic area—the geofence—
which is typically a circle with a specified radius. Pet.
App. 285a. Then it identifies a certain span of time, and
requests Location History data for all users within that
area during that time.
Pet. App. 285a.
Law
enforcement’s early geofence requests were broad,
asking Google to produce account-identified information
of all users caught in the geofence. Pet. App. 286a.
Concerned by the threat these requests posed to user
privacy, Google developed its own three-step response
process.
First, law enforcement must obtain a warrant
compelling Google to disclose a “de-identified” list of all
Google users whose Location History indicates they
were in the geofence at any point during the specific
time. Pet. App. 286a. Then, to create the list required
by the warrant, Google searches all Location History
user accounts in the Sensorvault for users whose stored
Location History indicates their location was
somewhere in the geofence. Pet. App. 287a. Google then
compiles a list for law enforcement including, for each
user, the stored latitude/longitude coordinates and
timestamp, the confidence interval (Google’s estimate of
how accurate the location data is), and the source of the
Location History (i.e., whether it came from Wi-Fi, GPS,
or cell tower). Pet. App. 287a-288a. At this step, each
user’s information is connected to a “de-identified”
unique device account number, rather than an email or
name. Pet. App. 287a.
10
Second, the government reviews the list of deidentified data. If needed, it may then ask Google for the
Location History of the users identified in step one for a
longer period of time than the original geofence, without
the geographic barrier. Pet. App. 289a-290a. This gives
the government a more complete picture of the relevant
users’ movements. Typically, Google requires the
government to narrow its request to a subset of accounts
from the original geofence. Pet. App. 290a.
Third and finally, the government compels Google to
provide the names and account identifiers—either email
addresses or phone numbers—associated with
particular devices it analyzed at step two. Thus, for
those users, the government obtains a complete and
granularly detailed record of their movements,
associated with their names and email addresses. While
Google prefers that law enforcement winnow down the
accounts it analyzed at step two, it is “possible” that
Google would provide account-identified location data
for all users in step two. Pet. App. 290a-291a.
In December 2023, Google announced that it would
begin saving users’ Location History on their devices,
rather than in the Sensorvault repository. See Letter
from Appellant under Fed. R. App. P. 28(j) (4th Cir. Dec.
20, 2023), Dkt. No. 62. Though Google announced that it
intended to make this change gradually over the year
between 2023 and 2024, petitioner is not aware of any
subsequent announcements that the change has been
made, detailing how the change affects legacy Google
devices, or specifying which data Google continues to
collect. Geofence warrants continue to be litigated. See
e.g. Stipulation Order to Partially Unseal Records in
11
Sealed Case, In re Google, Case No. 25-mj-70146 (N.D.
Cal. Feb. 7, 2025), ECF No. 1, Exhibit A (motion to
quash geofence warrant filed on February 7, 2024).
B. Factual Background
On May 20, 2019, someone robbed the Call Federal
Credit Union in Midlothian, Virginia. Pet. App. 265a266a. Upon entering the bank, the suspect handed the
teller a note demanding $100,000. Pet. App. 291a; 266a.
He then forced the manager to open the vault at
gunpoint, took $195,000 in cash, and left the bank on foot.
Pet. App. 266a-267a. Detective Hylton of the local police
responded to the scene, interviewed witnesses, and
reviewed surveillance video. Pet. App. 291a. After
unsuccessfully pursuing other leads, on June 14, 2019,
Detective Hylton applied for and obtained a geofence
warrant, ultimately leading to petitioner’s arrest. Pet.
App. 292-293a.
Detective Hylton followed Google’s three-step
procedure. He presented a magistrate judge with a
warrant for a geofence with a 300-meter diameter—
longer than three football fields—drawn over the site of
the robbery, encompassing a swath of urban Midlothian
and including both the bank and a nearby church. Pet.
App. 294a. At step one, the warrant called for deidentified information for all user accounts inside the
geographical area from 4:20 pm to 5:20 pm. Pet. App.
295a. For step two, it expanded the time frame from 3:50
to 5:50 and lifted the geographic limits. Pet. App. 296a.
And at step three, the warrant called for the accounts to
be linked to specific, identifiable users. Pet. App. 296a297a. In the warrant’s accompanying affidavit, Detective
Hylton stated that the perpetrator appeared to be using
12
a cell phone in the surveillance footage, and that “when
people act in concert with one another to commit a crime,
they frequently utilize cellular telephones.” Pet. App.
296a. The magistrate judge signed it, and Detective
Hylton sent it to Google on June 20, 2019. Pet. App.
298a.
Google executed step one of the geofence and
provided Detective Hylton with anonymized data for
nineteen users caught within the geofence. Pet. App.
299a.
Detective Hylton subsequently requested
additional location data on nine users, which Google
provided. Detective Hylton did not explain to Google
why he chose these nine accounts, nor did he consult a
magistrate. Pet. App. 299a-300a. Finally, Detective
Hylton requested that Google de-anonymize three of the
numbers, again without explaining why or consulting a
judge. Pet. App. 300a. Google provided the information.
Pet. App. 300a.
Ultimately, the geofence warrant led law
enforcement to petitioner, who was indicted on charges
of Forced Accompaniment During Armed Bank
Robbery, 18 U.S.C. § 2113(a), (d), and (e), and Using,
Carrying, or Brandishing a Firearm During and in
Relation to a Crime of Violence. 18 U.S.C. § 924(c)(1)(A).
C. District Court Proceedings
Petitioner moved to suppress the fruits of the
geofence warrant.
The district court conducted
extensive fact-finding, relying on, among other things,
an amicus brief by Google, four declarations by Google
employees, and in-person testimony from Google
employees. Pet. App. 306a-308a. In the end, the district
court concluded that the warrant violated the Fourth
13
Amendment because the government lacked “any
semblance of … particularized probable cause” to search
every one of the nineteen users swept into the geofence.
Pet. App. 312a. It further concluded that steps two and
three—undertaken with no judicial oversight—
“improperly provided law enforcement and Google with
unbridled discretion to decide which accounts will be
subject to further intrusions,” and independently failed
the Fourth Amendment’s particularity requirement.
Pet. App. 313a.
The court rejected the government’s argument that
petitioner forfeited his Fourth Amendment rights by
providing his Location History to Google. It concluded
that under Carpenter, “a user simply cannot forfeit the
protections of the Fourth Amendment for years of
precise location information by selecting ‘YES, I’M IN’
at midnight,” and that Google’s warnings were
insufficient to provide for voluntary consent. Pet. App.
333a. Finally, the court noted that geofence warrants
intrude into the private lives of many Americans who
may never have a chance to contest the incursion. Pet.
App. 310a. Nonetheless, the district court declined to
suppress the evidence through application of the goodfaith exception to the exclusionary rule. See United
States v. Leon, 468 U.S. 897, 923 (1984).
Petitioner entered a conditional guilty plea,
reserving his right to appeal the denial of his motion to
suppress. Pet. App. 80a. He was sentenced to 141
months’ imprisonment. Id.
D. Fourth Circuit Proceedings.
A divided Fourth Circuit panel affirmed on different
grounds. In the majority, Judges Richardson and
14
Wilkinson concluded that no search occurred, because
petitioner voluntarily exposed his information to Google.
Pet. App. 148a. Dissenting, Judge Wynn concluded that
under Carpenter, the geofence warrant here resulted in
a Fourth Amendment search. Pet. App. 187a.
The Fourth Circuit reheard the case en banc and
affirmed in a one-sentence per curiam opinion without
any reasoning. Pet. App. 4a. The court divided 7-7 on
whether a Fourth Amendment search had occurred,
with one judge declining to reach the issue. Of the seven
judges who found a Fourth Amendment search,
however, six concluded that the evidence should not be
suppressed under the good-faith exception to the
exclusionary rule, with Judge Gregory dissenting. Nine
judges wrote separate opinions airing “widely divergent
views.” Pet. App. 5a.
Seven judges—Judges Wilkinson, Niemeyer, King,
Agee, Richardson, Quattlebaum, and Rushing—found
that there was no Fourth Amendment search. Four
wrote separately.
Judge Richardson, joined by Judges Wilkinson,
Niemeyer, King, Agee, Quattlebaum, and Rushing,
concluded that no Fourth Amendment search occurred,
and therefore no warrant was necessary. Pet. App. 81a.
According to Judge Richardson, petitioner “did not have
a reasonable expectation of privacy in two hours’ worth
of Location History data voluntarily exposed to Google.”
Pet. App. 80a-81a. Judge Richardson distinguished
Carpenter on two grounds: first, Carpenter did not apply
to a request for only two hours of data, which he
characterized as encompassing only a “single, brief trip,”
and second, Carpenter did not apply to Location History
15
because users “knowingly and voluntarily” expose this
data to Google. Pet. App. 92a; 93a.
Judge Wilkinson, joined by Judges Niemeyer, King,
Agee, and Richardson, concluded that no search
occurred under the third-party doctrine, arguing that
the case involved a “straightforward” application of
Smith v. Maryland, 442 U.S. 735 (1979), and United
States. v. Miller, 425 U.S. 435 (1976). Pet. App. 23a. He
urged that treating collection of Location History as a
search would allow “tech-savvy criminals” to commit
crimes without consequence. Pet. App. 25a.
Judge Niemeyer opined that collecting Location
History is akin to collecting “markers” from the public
place of a crime. In Judge Niemeyer’s view, just as
police officers may collect boot prints, tire tracks, leftbehind items or DNA from a crime scene, so too may
they collect Location History through a geofence. Pet.
App. 33a.
Judge King noted his concurrence with Judges
Wilkinson and Richardson, and also stated that he would
affirm based on the good-faith exception. Pet. App. 36a.
Seven judges—Judges Gregory, Wynn, Thacker,
Harris, Heytens, Benjamin, and Berner—concluded that
a Fourth Amendment search had occurred. Two of those
judges (Judges Harris and Heytens) stopped there and
concluded that the good-faith exception to the
exclusionary rule applied; four judges (Judges Wynn,
Thacker, Benjamin, and Berner) went on to find that the
Fourth Amendment was violated, but also found the
good-faith exception applicable; and one judge (Judge
Gregory) would have excluded the evidence and
reversed petitioner’s conviction.
16
Judge Wynn, joined by Judges Thacker, Harris,
Benjamin, and Berner (in full) and by Judge Gregory
(except as to the good-faith exception), concluded that a
Fourth Amendment search had occurred under
Carpenter.
He emphasized that in light of the
“frictionless” nature of enabling Location History as
well as its “comprehensive, retrospective, intimate, and
highly efficient surveillance,” the third-party doctrine
was “wholly inadequate” to obviate petitioner’s
reasonable expectation of privacy in his Location
History data. Pet. App. 69a. Judge Wynn characterized
the en banc court as “squander[ing]” “a critical
opportunity to clarify the Fourth Amendment’s
application to emerging surveillance technologies.” Pet.
App. 38a. In declining to decide, Judge Wynn cautioned,
the court not only “clear[ed] the path for widespread,
surreptitious police surveillance,” but also fell “short of
[its] duty.” Pet. App. 70a-71a. He concluded, however,
that the evidence should not be excluded under the goodfaith exception to the exclusionary rule. Pet. App. 38a.
Judge Berner, joined by Judges Gregory, Wynn,
Thacker, and Benjamin, and in part by Judge Heytens,
similarly concluded that a search occurred. Judge
Berner interpreted Carpenter to lay out two factors
governing whether one holds a reasonable expectation
of privacy in information given to a third party: “(1) how
revealing that data is, and (2) whether the information
was, in practical terms, given to the third party
voluntarily.” Pet. App. 111a. She concluded that under
Carpenter, law enforcement conducts a search when it
obtains “any amount” of “non-anonymous” Location
History. Pet. App. 111a-112a. Judge Berner noted that
at Google’s first step, the warrant required only one
17
hour of information gleaned from primarily public
streets. Pet. App. 115a. Because of the request’s short
duration, limited geofence size, and coverage of only
public areas, she concluded that the data revealed in step
one was unlikely to be traceable to any specific
individual. Pet. App. 115a. At step two, however, the
government requested, and Google provided, nongeographically limited data from a longer period of time,
which likely would—and did—show users entering
homes and offices. Pet. App. 101a. The step two data,
then “was not truly anonymous.” Id. And at step three,
the data became explicitly non-anonymous, because the
government asked Google to reveal the names, email
addresses, and phone numbers associated with the
relevant Google users. Id. Steps two and three, Judge
Berner determined, infringed petitioner’s reasonable
expectation of privacy. Pet. App. 121a-122a, 125a.
In a portion of her opinion not joined by Judge
Heytens, Judge Berner concluded that the search
violated the Fourth Amendment because the
government lacked probable cause to search any specific
Google user at the time it applied for the warrant. Pet.
App. 126a.
She emphasized that “[b]ecause the
detective could not explain why he would eventually
search the Location History data of certain, thenunknown users in Google’s dataset, he failed to show
probable cause to conduct the second and third
requests.” Pet. App. 126a. “Under the terms of the
geofence warrant, Google, not a magistrate, was the sole
entity that could confine the scope of the ultimate
search.”
Pet. App. 126a.
“Probable cause
determinations cannot be delegated to private entities.”
Pet. App. 126a. Judge Berner disagreed, however, with
18
the Fifth Circuit’s view that geofence warrants are
always unconstitutional. Pet. App. 126a.
Judge Heytens, joined by Judges Harris and Berner,
wrote that regardless of whether a Fourth Amendment
violation occurred, exclusion of the evidence was
unwarranted because first, “the legal landscape was
uncertain when this investigation happened,” and
second, Detective Hylton sought the advice of
prosecutors and obtained a warrant. Pet. App. 98a.
Judge Gregory dissented from the per curiam
opinion. He concluded that the Fourth Amendment was
violated, the good-faith exception should not apply, and
the evidence should be suppressed. Judge Gregory
noted that though Detective Hylton obtained a warrant,
“[e]ven now, the government cannot tell us what
justified the more intrusive searches at [s]tep [t]wo or
[t]hree, or how or why there was probable cause to
search those individuals,” leaving it with “unbridled
discretion.” Pet. App. 133a, 135a. In Judge Gregory’s
view, to hold that lack of judicial guidance as to a specific
measure immunized it from exclusion “would run the
risk of forgiving law enforcement impropriety simply
because no court has specifically forbidden it,” rendering
Fourth Amendment protections “a nullity in the face of
rapidly emerging technology.” Pet. App. 139a.
Finally, Chief Judge Diaz, unlike his 14 colleagues,
declined to decide whether a search had occurred.
Instead, he voted to affirm the district court’s decision
solely through application of the good-faith exception.
Pet. App. 5a. He noted the divergence among his
colleagues: “[o]ne camp insists that disallowing geofence
warrants would contravene our precedent, hamstring
19
law enforcement…and chill innovation,” while the other
“is just as adamant that granting blanket approval to
these warrants would contravene our precedent and
compromise the privacy interests of cell phone users.”
Pet. App. 14a. But in light of the “shallow well of…legal
authority,” Judge Diaz chose to “wait.” Pet. App. 14a.
REASONS FOR GRANTING THE PETITION
I.
LOWER COURTS ARE DIVIDED ON THE
CONSTITUTIONALITY OF GEOFENCE
WARRANTS.
The Fourth Circuit’s judges were intractably divided
over the application of the Fourth Amendment to
geofence warrants.
That internal disagreement
parallels the disagreement among other courts.
To begin, a unanimous Fifth Circuit panel—in
conflict with the views of judges on both sides of the
Fourth Circuit’s divide—concluded that geofence
warrants always violate the Fourth Amendment.
United States v. Smith, 110 F.4th 817 (5th Cir. 2024),
petition for cert. filed, No. 24-7237 (U.S. May 19, 2025).
The Fifth Circuit first held that collecting Location
History is a Fourth Amendment search. In the Fifth
Circuit’s view, “[g]iven the intrusiveness and ubiquity,”
Google users have a reasonable expectation of privacy in
their Location History data. Id. at 836. The court
concluded that the third-party doctrine did not apply
under a “straightforward” application of Carpenter. Id.
at 835.
The Fifth Circuit further held that geofence
warrants are unconstitutional because they violate the
Fourth Amendment’s particularity requirement. It
20
explained that geofence warrants are a modern
incarnation of the “reviled ‘general warrants’ and ‘writs
of assistance’ of the colonial era.” Id. at 836. General
warrants specify an offense but allow officers the
discretion to determine which persons to arrest and
which places to search. Id. The Fifth Circuit drew a
parallel to the first step of Google’s geofence response
protocol, which is to search the entire Sensorvault
database—all 592 million individual accounts—for users
who were at a particular location at a given moment,
even though law enforcement does not yet know—and
may never know—who they are looking for. Id. at 837.
Accordingly, the Fifth Circuit held that “geofence
warrants fail at Step-1—they allow law enforcement to
rummage through troves of location data from hundreds
of millions of Google users without any description of the
particular suspect or suspects to be found.” Id. at 83738. The court held, however, that the evidence should
not be suppressed under the good-faith exception to the
exclusionary rule. Id. at 839.
The Texas Court of Criminal Appeals has rejected
the Fifth Circuit’s reasoning in Smith, resulting in a
conflict of authority between federal and state courts in
the same jurisdiction. In Wells v. State, No. PD 0669-23,
-- S.W.3d --, 2025 WL 980996 (Tex. Crim. App. Apr. 2,
2025), the court considered the constitutionality of a
warrant that followed Google’s three-step process. Like
the Fourth Circuit, the Texas Court of Criminal Appeals
fractured badly. Although no single opinion attracted a
majority of the court, eight of the nine Justices agreed
that steps one and two of Google’s three-step process
satisfied the Constitution, thus rejecting the Fifth
Circuit’s approach. See id. at *10 (opinion of Yeary, J.)
21
(declining to decide whether search occurred but finding
that warrant was supported by probable cause and was
sufficiently particularized); id. at *13 (opinion of Finley,
J.) (finding no search); id. at *16 (opinion of Newell, J.)
(finding no search).
As for step three, the Justices could not agree on the
appropriate disposition, with four Justices voting to
affirm the conviction based on varying rationales, id. at
*10 (opinion of Yeary, J.); id. at *13 (opinion of Finley,
J.), four Justices voting to vacate, id. at *18 (opinion of
Newell J.), and one Justice, Justice McClure, dissenting
without opinion.
The Georgia Supreme Court has similarly rejected
the Fifth Circuit’s Smith decision. In Jones v. State, 913
S.E.2d 700 (Ga. 2025), the court upheld the
constitutionality of a geofence warrant directed to
Google. The court did not decide whether a search had
occurred, id. at 706 n.1, instead holding that the warrant
satisfied the Fourth Amendment’s probable cause and
particularity requirements. The court concluded that
the geofence warrant was supported by probable cause
as to steps 1 and 2. Id. at 707. Contrary to the Fifth
Circuit, the Georgia Supreme Court determined that
“the description of the search in steps one and two of the
first warrant was not the kind that raises the specter of
a general rummaging, and it thus satisfies the
particularity requirement.” Id. at 711. Unlike in this
case, law enforcement in Jones obtained a second search
warrant at step three, and the court held that the second
search warrant was supported by probable cause. Id. at
708.
22
The Colorado Supreme Court, addressing a slightly
different issue, has also reached a conclusion
inconsistent with the Fifth Circuit. In People v.
Seymour, 536 P.3d 1260 (Colo. 2023), the Colorado
Supreme Court addressed the constitutionality of a
reverse keyword warrant, which required Google to
identify users who had done a particular search within a
particular period. Contrary to the Fifth Circuit, the
court held that the reverse keyword warrant was not a
general warrant and did not violate the Fourth
Amendment’s particularity requirement. Id. at 1276. It
reasoned that “a search isn’t unconstitutional simply
because the government, in some lightning-fast, digital
sense, very cursorily examines unrelated documents …
[s]uch brief examination doesn't turn an adequately
particularized search warrant into an unconstitutional
general warrant.” Id. The court ultimately did not
resolve the defendant’s other Fourth Amendment
objections and instead affirmed based on the good-faith
exception to the exclusionary rule. Id. at 1278.
Three members of the court would have held, like the
Fifth Circuit, that a warrant directing Google to search
through all accounts was an unconstitutional general
warrant. Justice Berkenkotter concurred based on the
good-faith exception but “strongly disagree[d] with the
majority’s conclusion that the examination of a billion
Google users’ search histories was not unreasonably
intrusive because the government didn’t ultimately
seize all of those search histories.” Id. at 1281
(Berkenkotter, J., concurring). The dissenting Justices
likewise concluded that “by authorizing law enforcement
to rummage through the private search histories of a
billion individuals for potential evidence of criminal
23
activity, reverse-keyword warrants permit exactly what
the Fourth Amendment forbids. They are tantamount to
a high-tech version of the reviled ‘general warrants’ that
first gave rise to the protections in the Fourth
Amendment.” Id. at 1282 (Marquez, J., dissenting).
II.
THIS CASE WARRANTS
COURT REVIEW.
SUPREME
The Court should grant certiorari in this case. The
question presented is profoundly important, the
confusion in lower courts is intolerable, and this case is
an appropriate vehicle given the well-developed facts
and well-preserved legal arguments.
A. The Issue is Important.
The question presented is both practically and
jurisprudentially important.
In recent years, law enforcement use of geofence
warrants has skyrocketed. Google received its first
geofence warrant in 2016. Pet. App. 285a. From 2017 to
2018, Google experienced a 1,500% increase in requests.
Pet. App. 285a. By 2021, geofence warrants constituted
25% of all warrants submitted to Google. Pet. App. 285a.
Geofence warrants have been used to investigate crimes
ranging from smashed windows to the storming of the
Capitol. United States v. Rhine, 652 F. Supp. 3d 38, 46
(D.D.C. 2023).
This fundamental new law enforcement tool has
attracted widespread attention. Numerous Fourth
24
Amendment scholars have analyzed the issue,1 with a
leading Fourth Amendment scholar observing that the
decision below yields “a crazy amount of uncertainty.”2
Given the ubiquity of location-tracking services,
geofence warrants have also resulted in substantial
public commentary. 3
1
See e.g. Brian L. Owsley, The Best Offense Is A Good Defense:
Fourth Amendment Implications of Geofence Warrants, 50 Hofstra
L. Rev. 829, 834 (2022); Haley Amster & Brett Diehl, Note, Against
Geofences, 74 Stan. L. Rev. 385, 389 & n.11 (2022); Note, Geofence
Warrants and the Fourth Amendment, 134 Harv. L. Rev. 2508,
2512 (2021); A. Reed McLeod, Note, Geofence Warrants:
Geolocating the Fourth Amendment, 30 Wm. & Mary Bill Rts. J.
531, 532 (2021); Orin S. Kerr, Data Scanning and the Fourth
Amendment (Stan. Pub. L., Working Paper, 2025),
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5175686; Orin
Kerr, The Fifth Circuit Shuts Down Geofence Warrants—and
Maybe
A
Lot
More,
Lawfare
(Aug.
14,
2024),
https://www.lawfaremedia.org/article/the-fifth-circuit-shuts-downgeofence-warrants-and-maybe-a-lot-more; Jackie O’Niel, Much
Ado About Geofence Warrants, Harv. L. Rev. Blog. (Feb. 18, 2025),
https://harvardlawreview.org/blog/2025/02/much-ado-about-geofen
ce-warrants/.
2
Orin S. Kerr, The Fourth Circuit’s Geofencing Case Ends Not
With a Bang But A Whimper, reason (May 2, 2025),
https://reason.com/volokh/2025/05/02/the-fourth-circuits-geofenc
ing-case-ends-not-with-a-bang-but-a-whimper/.
3
See e.g. Andrew Couts, Security News This Week: Geofence
Warrants Ruled Unconstitutional—but That’s Not the End of it,
WIRED (Aug. 17, 2024), https://www.wired.com/story/geofencewarrants-ruled-unconstitutional-tmobile-fine-deepfake-nudes-law
suit/; Shira Ovide, Police Love Google’s Surveillance Data: Here’s
How to Protect Yourself, Wash. Post (Oct. 24, 2023),
https://www.washingtonpost.com/technology/2023/10/24/google-pri
vacy-police-geofence/; Sidney Fussell, An Explosion in Geofence
Warrants Threatens Privacy Across the U.S., Wired (Aug. 27,
25
Yet tech companies and magistrate judges have been
forced to grapple with these difficult and consequential
questions on their own. Tech companies have had no
choice but to develop protocols, without judicial
guidance, for balancing law enforcement interests with
user privacy. The three-step procedure in this case was
designed by Google.
This Court—not a private
business—should decide how the Fourth Amendment
works in the context of geofence warrants.
What is more, magistrate judges must reinvent the
wheel every time they are faced with a geofence
warrant. For ordinary warrants, a magistrate judge’s
duties are well-understood. The magistrate judge must
assess whether there is probable cause and the warrant
is sufficiently particularized—tasks magistrate judges
accomplish every day.
Geofence warrants, by contrast, are a fundamentally
new and different type of warrant, forcing magistrate
judges to engage in first-principles constitutional
analysis. And that analysis has not proved easy. As
catalogued above, judicial views of geofence warrants
have spanned the gamut, ranging from the view of seven
Fourth Circuit judges that geofence warrants are never
needed because collection of location information is not a
search, to the Fifth Circuit’s view that geofence
warrants result in the unconstitutional search of
hundreds of millions of private user accounts every time
they are executed. The Fourth Circuit and Texas Court
2021),
https://www.wired.com/story/geofence-warrants-google/;
Jennifer Valentino-deVries, Tracking Phones, Google is a Dragnet
for the Police, N.Y. Times (Apr. 13, 2019), https://www.nytimes.
com/interactive/2019/04/13/us/google-location-tracking-police.html.
26
of Criminal Appeals produced nearly as many
intermediary positions as there are judges. Magistrate
judges should not be forced to evaluate and pick one of
these positions every time they receive a warrant
application. It is incumbent on this Court to decide these
fundamental constitutional questions.
Certiorari is warranted, too, in view of the privacy
implications of geofence warrants. As this Court has
recognized, cellphones are now so “pervasive and
insistent” in modern life that to an extraterrestrial
visitor, they might be mistaken for a “feature of human
anatomy.” Riley v. California, 573 U.S. 373, 385 (2014).
Most Americans “compulsively” carry their phone at all
times. Carpenter, 585 U.S. at 311. The record in this
case reflects that there are 1.5 billion Google users, of
which about one third—hundreds of millions—have
enabled Location History. Pet. App. 274a. As with the
CSLI this Court addressed in Carpenter, those phones’
location-tracking capabilities produce a “deep repository
of historical location information” of users’ daily
movements accessible with “just the click of a button”
and at “practically no expense.” Carpenter, 585 U.S. at
311-12.
And just as in Carpenter, because this
information is collected for nearly all users at all times,
“police need not even know in advance whether they
want to follow a particular individual, or when.” Id. at
312. The resulting system of “near perfect surveillance”
is as if the government had “attached an ankle monitor”
to tens of millions of Americans. Id.
Under the Fifth Circuit’s view of geofence warrants,
the government rummages through every user’s private
location information every time it executes a geofence
27
warrant. If that holding is right, millions of Americans
are having their privacy rights violated every day. Even
if that holding is wrong, millions of Americans face the
risk that their identities will be unmasked and
movements revealed without any judicial involvement.
If the Constitution permits that outcome, this Court
should say so.
B. This Case Is an Appropriate Vehicle.
The Court is unlikely to find a better vehicle than this
case to resolve the constitutionality of geofence
warrants.
First, the Fourth Circuit’s one-line per curiam
affirmance leaves all issues on the table for this Court.
The Court is free to resolve whichever issues it deems
appropriate—whether a search occurred, whether the
warrant was sufficiently particularized, whether there
was probable cause, or anything else—and leave other
issues for the Fourth Circuit on remand.
Second, petitioner preserved all available Fourth
Amendment arguments, and those arguments were the
subject of widespread debate in the 126 pages of Fourth
Circuit opinions. As Judge Wynn observed, “[t]he
constitutional question in this case has been fully
briefed, argued and exhaustively debated—not only by
the parties but by amici and members of this Court. And
it is unclear what future case could better tee up the
issue.” Pet. App. 38a.
Of particular note, petitioner preserved the
argument that he had a property right in Location
History. In Carpenter, Justice Gorsuch suggested that
the Fourth Amendment analysis may have turned on
28
whether users had a property interest in CSLI, but
could not “help but conclude—reluctantly—that Mr.
Carpenter forfeited perhaps his most promising line of
argument.” 585 U.S. at 406 (Gorsuch, J., dissenting). In
this case, by contrast, in the district court, petitioner
repeatedly briefed the issue that he had a property
interest in CSLI. 4th Cir. J.A. 39-40; J.A. 83-85; J.A. 382383; J.A. 1102; J.A. 1172-1173.
The government
expressly acknowledged that petitioner had preserved a
property-based theory “rooted in Justice Gorsuch’s solo
dissent in Carpenter.” J.A. 62. In the Fourth Circuit,
petitioner similarly argued that “[r]egardless of the
duration of the search or the significant privacy
interests at stake, Location History data fits into a
simpler scheme: the Fourth Amendment protects it
because it belongs to the users who created it.” 4th Cir.
Reply Br. 11.
The factual record in this case is also unusually
extensive. Google participated in the district court as an
amicus, providing detailed information about the
technology involved as well as its internal processes. See
Dkt. No. 29, Brief of Amicus Curiae Google LLC in
Support of Neither Party Concerning Defendant’s
Motion to Suppress Evidence from a ‘Geofence’ General
Warrant, 2019 WL 8227162. A Google “Location History
Manager” submitted three declarations, a Google “Legal
Investigations Specialist” submitted another, and both
testified live. Indeed, the Fifth Circuit in Smith often
relied on the findings of the district court from Chatrie
for its description of what Location History is, how
precise it is, how it is enabled, Google’s geofence warrant
protocol, and the ubiquity of geofences. See e.g., 110
F.4th at 823-25. In short, the Court is unlikely to see
29
another case in which both the facts and the law are so
well-developed.
III.
PETITIONER’S FOURTH AMENDMENT
RIGHT WAS VIOLATED.
The Court should hold that the geofence warrant in
this case violated the Fourth Amendment.
A. The Government Conducted a Search.
When the government directed Google to execute
the geofence warrant, it conducted a search under both
the Carpenter majority’s reasonable-expectations-ofprivacy approach and under a property-based approach.
First, the government violated petitioner’s
reasonable expectation of privacy.
The Fourth
Amendment protects individuals against “arbitrary
invasions by governmental officials” into the sphere in
which a person has a reasonable expectation of privacy.
Carpenter, 585 U.S. at 303 (quoting Camara v. Mun. Ct.
of City & Cnty. of S.F., 387 U.S. 523, 528 (1967)).
Accordingly, when an individual “seeks to preserve
something as private,” and his “expectation of privacy is
one that society is prepared to recognize as reasonable,”
intrusion into that sphere is a search. Id. at 304 (quoting
Katz v. United States, 389 U.S. 347, 351 (1967)).
The Carpenter Court concluded that cell phone users
have a reasonable expectation of privacy in their CSLI,
and the same reasoning establishes that they have a
reasonable expectation of privacy in their Location
History. The Court recognized that CSLI is data that a
person would reasonably expect to keep private, as it
tracks “nearly exactly the movements of [a cell phone’s]
owner,” id. at 311, allows the police to access
30
“retrospective” data, id. at 312, provides an “intimate
window into a person’s life,” id., and is “easy, cheap, and
efficient.” Id. at 311.
As Judge Wynn and Judge Berner explained in
detail, and as the Fifth Circuit explained in Smith, each
of these characteristics of CSLI is just as true of
Location History data. Pet. App. 62a; Pet. App. 118a;
Smith, 110 F.4th at 832-33. Location History offers even
more precise tracking than CSLI and provides a detailed
map of all of a person’s movements, including movement
into the most intimate of spaces. See Carpenter, 585 U.S.
at 311 (noting that a cell phone “faithfully follows its
owner…into private residences, doctor’s offices, political
headquarters, and other” revealing locales). It is also
easy and cheap to obtain.
Carpenter also establishes that petitioner did not
waive the protections of the Fourth Amendment by
transmitting his Location History to Google.
In
Carpenter, the Court rejected the government’s proffer
to treat CSLI as simply a “garden-variety request for
information from a third-party witness.” Id. at 313.
First, CSLI was intimately revealing, unlike the pen
register and bank records at issue in Smith and Miller,
the foundational third-party cases.
And second,
cellphone users do not, in any “meaningful sense,” opt in
to “turning over a comprehensive dossier of [their]
physical movements.” Id. at 315.
Transmission of Location History to Google is not
meaningfully voluntary. As Judge Wynn and the Fifth
Circuit explained, enabling Location History is
“frictionless by design” and difficult to undo. Pet. App.
65a; see Smith, 110 F.4th at 835. Meanwhile, Google’s
31
privacy warnings are “limited” and “partially hidden,”
and at the very least, do not inform users that enabling
the feature might result in the government’s unfettered
access to their every move. Pet. App. 66a, 332a; Smith,
110 F.4th at 835-36.
As a result, the execution of the geofence warrant
resulted in a search. As the Fifth Circuit explained, at
the first step, the government searched the accounts of
all users by directing Google to assess whether those
users were within the geofence. See Smith, 110 F.4th at
837-38 (geofence warrants “allow law enforcement to
rummage through troves of location data from hundreds
of millions of Google users without any description of the
particular suspect or suspects to be found”). Although
this search did not unmask any particular user’s identity,
it was still a search—just as the police conducts a search
if it rummages through a person’s desk, even if it does
not know who owns the desk. At a minimum—in line
with Judge Berner’s conclusion—the government
conducted a search when it obtained Chatrie’s accountidentified location information. Pet. App. 121a. Google’s
disclosure of non-anonymized information regarding
petitioner’s
whereabouts
violated
petitioner’s
reasonable expectation of privacy.
Under a property-based approach, the government
conducted a search because Location History belongs to
users, not Google. Location History is their digital
papers and effects, their personal “journal” stored in
their accounts, just like their Gmail, Google Docs, or
Google Photos. Google’s privacy policy consistently
refers to user data (including Location History) as “your
information,” which can be managed, exported, and even
32
deleted from Google’s servers at “your” request. 4th
Cir. J.A. 39. Unlike the CSLI at issue in Carpenter,
Location History is in no sense a business record: users
(not Google) have control over whether it is stored and
can unilaterally delete it. Although the records were
stored on Google’s servers, Google was acting as a bailee
rather than an owner. Having directed Google to screen
and then turn over the contents of petitioners’ property,
the government conducted a search.
B. The
Warrant
Amendment.
Violated
the
Fourth
Although the government obtained a geofence
warrant before conducting the search, the warrant did
not comply with the Fourth Amendment.
First, the warrant violated the Fourth Amendment’s
particularity requirement. To comply with the Fourth
Amendment, warrants must “particularly describ[e] the
place to be searched, and the persons or things to be
seized.” U.S. Const. amend IV. As the Fifth Circuit’s
Smith decision explains, geofence warrants are not
particularized. “[L]aw enforcement cannot obtain its
requested location data unless Google searches through
the entirety of its Sensorvault—all 592 million individual
accounts—for all of their locations at a given point in
time.” 110 F.4th at 837. “Moreover, this search is
occurring while law enforcement officials have no
idea who they are looking for, or whether the search will
even turn up a result.” Id. “Indeed, the quintessential
problem with these warrants is that they never include
a specific user to be identified, only a temporal and
geographic location where any given user may turn up
post-search.” Id.
33
Even if geofence warrants are not categorically
unconstitutional, this warrant was. The district court
opined that it may be possible to narrowly describe the
circumference of a geofence that would be supported by
probable cause for the limited, anonymous information it
provides, from which “officers likely could use that
narrow, anonymous information to develop probable
cause particularized to specific users.” Pet. App. 325a.
But that is not what occurred here. Detective Hylton
sought and obtained a geofence warrant covering an
area of downtown Midlothian the size of three football
fields, with a confidence interval reach even larger.
Then, at step two, Detective Hylton offered no
reasoning to justify a more intrusive search of nine of the
nineteen—unlimited as to geographic area, and longer in
time—nor did he present his request for a more
intrusive search to a neutral judge or magistrate.
At the third step, when Detective Hylton directed
Google to unmask the identities of three account holders,
he conducted a warrantless search without probable
cause. Detective Hylton lacked probable cause to
investigate those three accounts: he offered no
reasoning to Google as to why these three accounts were
singled out. And even if Detective Hylton offered such
reasoning, Google is not a magistrate judge. Detective
Hylton should have obtained a search warrant,
supported by probable cause, before determining the
identities associated with the devices he found
suspicious. Because he failed to do so, the search
violated the Fourth Amendment.
IV.
34
THE GOOD-FAITH EXCEPTION DOES
NOT APPLY.
The first question—whether this geofence warrant
violated the Fourth Amendment—alone justifies this
Court’s intervention. Because the Fourth Circuit
resolved this case via an unreasoned per curiam
affirmance, the Court has the option of granting
certiorari on the first question, and then, if it finds that
an unconstitutional search occurred, remanding for the
Fourth Circuit to decide whether the good-faith
exception applies. Although the Fourth Circuit judges
have already opined on that issue in their separate
opinions below, they may reconsider their reasoning in
light of this Court’s Fourth Amendment analysis.
Further, deciding the Fourth Amendment merits issue
will guide lower courts and law enforcement on this
important constitutional issue.
Nevertheless, to ensure that petitioner obtains
meaningful relief from a decision in his favor, petitioner
respectfully urges the Court to grant certiorari as to the
second question presented and hold that the good-faith
exception to the exclusionary rule does not apply.
To begin, petitioner agrees with Judge Gregory’s
conclusion that the good faith exception does not apply
because “Hylton could not have reasonably believed that
the liberty authorized by the warrant was constitutional
given the lack of specificity the Fourth Amendment
explicitly demands.” Pet. App. 136a. Further, this
Court has never held that the good-faith exception
applies when the warrant’s defect is that it is an
unconstitutional general warrant—as opposed to lacking
adequate probable cause. Cf. Groh v. Ramirez, 540 U.S.
35
551, 563-64 (2004) (denying qualified immunity when
warrant plainly failed particularization requirement).
The good-faith exception, moreover, should not be
extended to this case because it will impede this Court
from ever deciding the important constitutional
questions presented here.
In the context of a typical search warrant, whether
the good faith exception applies or not turns on
questions—the amount of information supporting a
magistrate’s probable cause determination, and the
objective reasonability of an officer’s reliance on that
determination—that vary from case to case. Every
geofence warrant, by contrast, is issued based on the
same basic set of information: the existence of a crime, a
place, and a time, without any particular information as
to any particular suspect.
If the good-faith exception applies to geofence
warrants based upon that minimum quantum of
information, there would be no possible way for a ruling
to benefit the person that was searched, regardless of
procedural posture. In every civil case, officers will
assert qualified immunity based on the current
uncertainty in the law. In every criminal case, officers
will argue that the very fact that the geofence warrant
issued is a sufficient basis to apply the good-faith
exception. Indeed, in Smith, the Fifth Circuit accepted
that very argument. Hence, if the good-faith exception
applies, any opinion issued by this Court would be
advisory. See Kerr, supra note 4 (noting, in commentary
on decision below, that applying the good-faith exception
will result in advisory opinions in every case).
36
This problem did not arise in Leon. In Leon, the
question was whether evidence should be excluded that
was “obtained by officers acting in reasonable reliance
on a search warrant issued by a detached and neutral
magistrate but ultimately found to be unsupported by
probable cause.” 468 U.S. at 900. But questions about
probable cause can be resolved in cases not involving
warrants, such as vehicle search cases. See generally
Wyoming v. Houghton, 526 U.S. 295, 300-01 (1999). By
contrast, geofence warrant cases always involve
warrants based upon the same basic information that
vary only in the type of crime and the temporal and
geographic scope of the warrant obtained. The disputes
turn on whether the warrants are sufficiently
particularized and whether a new warrant is needed at
the final step before de-anonymizing data. Thus, in every
case, the government will invoke the good-faith
exception merely by virtue of having obtained a
warrant, thus preventing any defendant from benefiting
from
a
ruling
finding
geofence
warrants
unconstitutional.
That result is untenable. Geofence warrants are a
consequential and controversial new type of warrant.
This Court should be able to decide whether they are
constitutional in a case where its decision actually
matters.
This Court’s precedents suggest it may relax the
good-faith exception under these unusual circumstances.
In Davis v. United States, 564 U.S. 229 (2011), this Court
held that the good-faith exception applies when an
officer relies on case law that is subsequently overruled.
The defendant raised the concern that no defendant
37
could ever benefit from a Supreme Court decision
overruling precedent, impeding the law’s development.
In view of that concern, this Court recognized that “in a
future case, we could, if necessary, recognize a limited
exception to the good-faith exception for a defendant
who obtains a judgment overruling one of our Fourth
Amendment precedents.” Davis, 564 U.S. at 248. For
similar reasons, the Court should recognize a limited
exception to the good-faith exception for a defendant
who challenges the legality of a fundamentally new type
of warrant issued without any particularized
information about any suspect.
CONCLUSION
For the foregoing reasons, this Court should
grant the petition.
Respectfully submitted,
GEREMY C. KAMENS
Federal Public Defender
LAURA J. KOENIG
Assistant Federal Public
Defender
PATRICK BRYANT
Assistant Federal Public
Defender
OFFICE OF THE FEDERAL
PUBLIC DEFENDER,
EASTERN DISTRICT OF
VIRGINIA
1650 King Street,
Suite 500
Alexandria, VA 22314
(703) 600-0800
ADAM G. UNIKOWSKY
Counsel of Record
LAUREL A. RAYMOND
JENNER & BLOCK LLP
1099 New York Ave., NW
Suite 900
Washington, DC 20001
(202) 639-6000
AUnikowsky@jenner.com
APPENDIX
TABLE OF CONTENTS
Appendix A
United States v. Chatrie, 136 F.4th 100 (4th
Cir. 2025) (on rehearing en banc) ............................. 1a
Appendix B
United States v. Chatrie, No. 22-4489, 2024
WL 4648102 (4th Cir. Nov. 1, 2024) (order
granting rehearing en banc) ................................. 143a
Appendix C
United States v. Chatrie, 107 F.4th 319 (4th
Cir. 2024) .................................................................. 145a
Appendix D
United States v. Chatrie, 590 F.Supp.3d 901
(E.D. Va. 2022) ........................................................ 264a
1a
Appendix A
ON REHEARING EN BANC
UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT
____________________
No. 22-4489
____________________
UNITED STATES of America,
Plaintiff – Appellee,
v.
Okello T. CHATRIE,
Defendant – Appellant.
The Reporters Committee for Freedom of the Press;
American Civil Liberties Union; American Civil
Liberties Union of Virginia; Eight Federal Public
Defender Offices Within the Fourth Circuit;
Technology Law and Policy Clinic at New York
University School of Law; Electronic Frontier
Foundation,
Amici Supporting Appellant.
Project for Privacy and Surveillance Accountability,
Inc.,
Amicus Supporting Rehearing Petition.
Argued: January 30, 2025
Decided: April 30, 2025
2a
Appeal from the United States District Court for the
Eastern District of Virginia, at Richmond. M. Hannah
Lauck, District Judge. (3:19−cr−00130−MHL−1)
Before DIAZ, Chief Judge, and WILKINSON,
NIEMEYER, KING, GREGORY, AGEE, WYNN,
THACKER,
HARRIS,
RICHARDSON,
QUATTLEBAUM,
RUSHING,
HEYTENS,
BENJAMIN, and BERNER, Circuit Judges.
Affirmed by published per curiam opinion in which Chief
Judge Diaz, Judge Wilkinson, Judge Niemeyer, Judge
King, Judge Agee, Judge Wynn, Judge Thacker, Judge
Harris, Judge Richardson, Judge Quattlebaum, Judge
Rushing, Judge Heytens, Judge Benjamin, and Judge
Berner joined.
Chief Judge Diaz wrote a concurring opinion. Judge
Wilkinson wrote a concurring opinion, in which Judge
Niemeyer, Judge King, Judge Agee, and Judge
Richardson joined. Judge Niemeyer wrote a concurring
opinion. Judge King wrote a concurring opinion. Judge
Wynn wrote a concurring opinion, in which Judge
Thacker, Judge Harris, Judge Benjamin, and Judge
Berner joined in full, and in which Judge Gregory joined
except as to footnote 1. Judge Richardson wrote a
concurring opinion, in which Judge Wilkinson, Judge
Niemeyer, Judge King, Judge Agee, Judge
Quattlebaum, and Judge Rushing joined. Judge Heytens
wrote a concurring opinion, in which Judge Harris and
Judge Berner joined. Judge Berner wrote a concurring
opinion, in which Judge Gregory, Judge Wynn, Judge
Thacker, and Judge Benjamin joined in full, and in which
Judge Heytens joined as to Parts I, II(A), and II(B).
3a
Judge Gregory wrote a dissenting opinion.
ARGUED: Michael William Price, NATIONAL
ASSOCIATION
OF
CRIMINAL
DEFENSE
LAWYERS, Washington, D.C., for Appellant. Nathan
Paul Judish, UNITED STATES DEPARTMENT OF
JUSTICE, Washington, D.C., for Appellee. ON BRIEF:
Geremy C. Kamens, Federal Public Defender,
Alexandria, Virginia, Laura J. Koenig, Assistant
Federal Public Defender, OFFICE OF THE
FEDERAL PUBLIC DEFENDER, Richmond,
Virginia, for Appellant. Kenneth A. Polite, Jr., Assistant
Attorney General, Richard W. Downing, Deputy
Assistant Attorney General, UNITED STATES
DEPARTMENT OF JUSTICE, Washington, D.C.;
Jessica D. Aber, United States Attorney, Kenneth R.
Simon, Jr., Assistant United States Attorney, Peter S.
Duffey, Assistant United States Attorney, OFFICE OF
THE UNITED STATES ATTORNEY, Richmond,
Virginia, for Appellee. Jennifer Lynch, Andrew
Crocker, Hannah Zhao, ELECTRONIC FRONTIER
FOUNDATION, San Francisco, California; Jacob M.
Karr, Technology Law and Policy Clinic, NEW YORK
UNIVERSITY SCHOOL OF LAW, New York, New
York, for Amici Technology Law and Policy Clinic at
New York University School of Law. Jennifer Stisa
Granick, San Francisco, California, Nathan Freed
Wessler, Ashley Gorski, Patrick Toomey, Brandon
Buskey, Trisha Trigilio, Laura Moraff, Brett Max
Kaufman, AMERICAN CIVIL LIBERTIES UNION
FOUNDATION, New York, New York; Eden B.
Heilman, Matthew W. Callahan, AMERICAN CIVIL
LIBERTIES UNION FOUNDATION OF VIRGINIA,
Richmond, Virginia; William F. Nettles, IV, Federal
4a
Public Defender, Columbia, South Carolina, G. Alan
Dubois, Federal Public Defender, Raleigh, North
Carolina, Louis Allen, Federal Public Defender,
Greensboro, North Carolina, Juval O. Scott, Federal
Public Defender, Roanoke, Virginia, Brian J. Kornbrath,
Federal Public Defender, Clarksburg, West Virginia,
John Baker, Federal Public Defender, Charlotte, North
Carolina, James Wyda, Federal Public Defender,
Baltimore, Maryland, Wesley P. Page, Federal Public
Defender, OFFICE OF THE FEDERAL PUBLIC
DEFENDER, Charleston, West Virginia, for Amici
American Civil Liberties Union, American Civil
Liberties Union of Virginia, and Eight Federal Public
Defender Offices Within the Fourth Circuit. Bruce D.
Brown, Katie Townsend, Gabe Rottman, Grayson Clary,
Emily Hockett, REPORTERS COMMITTEE FOR
FREEDOM OF THE PRESS, Washington, D.C., for
Amicus The Reporters Committee for Freedom of the
Press. Gene C. Schaerr, Erik S. Jaffe, Aaron C. Ward,
SCHAERR | JAFFE LLP, Washington, D.C., for
Amicus Project for Privacy & Surveillance
Accountability, Inc.
PER CURIAM:
The judgment of the district court is
AFFIRMED.
5a
DIAZ, Chief Judge, concurring:
I join in affirming the district court’s denial of Okello
Chatrie’s suppression motion, but solely on the court’s
finding of good faith. See United States v. Chatrie, 590 F.
Supp. 3d 901, 936–41 (E.D. Va. 2022). My colleagues have
widely divergent views on the intersection of the Fourth
Amendment and the groundbreaking investigative tool
at issue here. I respect the care and attention they’ve
devoted to this matter. But judicial modesty sometimes
counsels that we not make grand constitutional
pronouncements merely because we can.
This is such a case.
I.
A.
Today we consider the constitutionality of geofence
warrants, a novel and powerful technology that law
enforcement has increasingly used to investigate crime.
In simple terms, a geofence warrant requires a service
provider to produce location data from cell phone users
who were near the scene when a crime occurred.
Like a traditional warrant, law enforcement (as here)
may apply for a geofence warrant from a judge. If
granted, law enforcement can then serve the warrant on
the provider (here, Google).1
1
The district court explained: “Other companies such as Amazon
and Apple invariably retain users’ location data as well. But Google,
whose services function across Apple and Android devices ..., seems
6a
Google collects the Location History of over 500 million
users, and it’s this data that law enforcement accesses
via a geofence warrant. Location History “appears to be
the most sweeping, granular, and comprehensive tool—
to a significant degree—when it comes to collecting and
storing location data.” Chatrie, 590 F. Supp. 3d at 907
(emphasis omitted).
It’s also remarkably extensive, “log[ging] a device’s
location, on average, every two minutes,” even “in terms
of elevation.” Id. at 908. If a device is in a building, for
example, its Location History can show on which floor.
When presented with a geofence warrant, Google
applies an internally developed three-step process,
providing to law enforcement an anonymous “list of all
Google users whose Location History data indicates
were within the geofence during a specified timeframe.”
Id. at 915 (cleaned up). To do this, “Google must search
all Location History data to identify users,” regardless
of whether the users “saved Location History data.” Id.
(cleaned up).2 After narrowing the list to users who had
their Location History enabled, Google also provides
“the date and time, the latitude and longitude, the
to be subject to more geofence requests than other companies.”
Chatrie, 590 F. Supp. 3d at 907 n.8. What’s more, “[c]ompanies such
as Apple, Lyft, Snapchat, and Uber have all received geofence
warrant requests, but Google is the most common recipient and ‘the
only one known to respond.’” United States v. Smith, 110 F.4th 817,
821 n.2 (5th Cir. 2024) (cleaned up).
2
Location History “is off by default” on a cell phone, though it’s
“‘possible that a user would have seen the option’ to opt into
Location History multiple times across multiple apps.” Id. at 908–
09.
7a
geolocation source used, and the map display radius (i.e.,
the confidence interval)” for the relevant accounts.
United States v. Smith, 110 F.4th 817, 824–25 (5th Cir.
2025).
At the second step, law enforcement may “compel
Google to provide additional location coordinates beyond
the time and geographic scope of the original request,”
ostensibly to “assist ... in eliminating devices.” Chatrie,
590 F. Supp. 3d at 916 (cleaned up). But while law
enforcement may widen the geographic scope of the
request, Google “typically require[s] law enforcement to
narrow the number of users for which it requests
[additional] data.” Id.
Finally, at the third step, law enforcement “‘can compel
Google to provide account-identifying information’ for
the users ‘the [g]overnment determines are relevant to
the investigation.’” Id. (cleaned up). “This ‘accountidentifying information’ includes the name and email
address associated with [an] account.” Id.
B.
The police charged Chatrie with two crimes related to a
bank robbery based on information obtained from
Google through a geofence warrant. Detective Joshua
Hylton prepared the warrant, which “drew a geofence
with a 150-meter radius—with a diameter of 300 meters,
longer than three football fields—in an urban
environment.” Id. at 918. That radius included the bank
and a nearby church. Id. The warrant “sought location
data for every device present within the geofence” for
8a
an hour around the time of the robbery (i.e., thirty
minutes before and thirty minutes after). Id. at 919.
In the warrant, Detective Hylton described Google’s
three-step process, explaining that he would “‘attempt
to narrow down’ the list of users for which the
[g]overnment would obtain the most invasive
information.” Id.
First, the warrant directed Google to “‘provide
“anonymized information” regarding the Accounts that
are associated with a device that was inside the
described geographical area’” in the hour around the
robbery. Id. Next, “[l]aw enforcement would return a
list of accounts that they had attempted to narrow
down,” so that “Google would then ‘produce contextual
data points with points of travel outside of the
geographical area.” Id. (cleaned up). To do so, “the
warrant expanded the timeframe to include thirty
minutes before and thirty minutes after the initial hourlong window”—covering a two-hour total window. Id.
Finally, law enforcement would direct Google to provide
identifying information for certain accounts.
In his affidavit supporting the warrant, Hylton added
that the geofence process could identify not only the
robber but also “potential witnesses and/or [other]
suspects.” Id. at 920. This was because the detective had
observed on surveillance footage that the robber “had a
cell phone in his right hand and appeared to be speaking
with someone else on the device”—someone with whom
the robber may have been “act[ing] in concert.” Id.
Using the warrant and the subsequent information
9a
Google provided, law enforcement identified Chatrie as
a suspect.
After his arrest, Chatrie, who had opted to share his
Location History with Google, moved to suppress the
location information, arguing that the warrant violated
the Fourth Amendment. The district court agreed that
this geofence warrant “plainly violate[d]” the
Constitution, 3 id. at 905, but nonetheless declined to
suppress it under the good-faith exception to the Fourth
Amendment, id. at 936–41.
The district court emphasized that “evidence obtained
pursuant to a search warrant issued by a neutral
magistrate need not be excluded if the officer’s reliance
on the warrant was ‘objectively reasonable.’” Id. at 937
(cleaned up). Ticking through the factors the Supreme
Court outlined in United States v. Leon, 468 U.S. 897, 104
S.Ct. 3405, 82 L.Ed.2d 677 (1984), that we have since
applied, see, e.g., United States v. Doyle, 650 F.3d 460,
467 (4th Cir. 2011), the district court found that the
instant warrant passed the good-faith bar. Chatrie, 590
F. Supp. 3d at 937.
When Detective Hylton applied for the geofence
warrant in this case, no court had ruled on the legality of
such warrants generally. So he relied on his experience,
3
The Fifth Circuit has held “that geofence warrants are general
warrants categorically prohibited by the Fourth Amendment.”
Smith, 110 F.4th at 838. But like the district court here, the Fifth
Circuit in Smith declined to suppress the challenged warrant on
good-faith grounds. Id. at 838–40.
10a
having successfully obtained three other geofence
warrants after consulting with prosecutors before
seeking them. Id. at 938.
Hylton also obtained approval from a state magistrate
for the warrant. See id. at 938–39. To be sure, neither the
detective nor the magistrate performed their duties
perfectly.
Inexplicably, Detective Hylton submitted a search
warrant return—which “notifies the Court when an
officer executes a search warrant” and describes “what
items [the officer] gathered during the search”—to the
magistrate before he had even served the warrant on
Google. Id. at 920. In that return, Hylton “stated that he
had executed the warrant,” even though, again, he
hadn’t yet sent it to Google. Id. And he wrote that he had
seized “Data,” when, in fact, he seized “what would be a
sizable amount of precise location information on at least
nineteen device users.” Id. (cleaned up).
As for the magistrate, he “asked no questions” of
Detective Hylton. Nor did he “seek to modify anything”
in the accompanying affidavit, even though this appears
to be the first geofence warrant application the
magistrate had considered. Id. at 918.
Still, the district court was satisfied that the warrant
was “not so lacking in indicia of probable cause as to
render official belief in its existence entirely
unreasonable.” Id. at 937 (cleaned up). The good-faith
exception thus saved the warrant from suppression.
I would adopt that narrow holding here.
11a
II.
A.
Geofence warrants are an extraordinary investigatory
advancement, born out of technological developments
enabling the relentless collection of eerily precise
location data. But questions remain about the
technology enabling such warrants as well as Google’s
process for responding to them. It’s no mystery then
that applying our legal precedents to this rapidly
evolving technology is precarious. Indeed, as the district
court noted, “[t]his case implicates the next phase in the
courts’ ongoing efforts to apply the tenets underlying
the Fourth Amendment to previously unimaginable
investigatory methods.” Chatrie, 590 F. Supp. 3d at 905.
Earlier cases applied the Fourth Amendment to
“recording devices in public telephone booths,”
“thermal-imaging equipment” aimed at homes, “and,
most recently, to cell-site location data.” Id.
(summarizing cases). The cases have protected “data
that provides an intimate window into a person’s life,
revealing not only his particular movements, but
through them his ‘familial, political, professional,
religious, and sexual associations,’” if that data hasn’t
been meaningfully disclosed to a third party. Carpenter
v. United States, 585 U.S. 296, 311, 314–15, 138 S.Ct.
2206, 201 L.Ed.2d 507 (2018).4
4
The Court opined that whether and how the Fourth Amendment
applied to cell-site records existed “at the intersection of two lines
of cases, both of which inform[ed] [its] understanding of the privacy
interests at stake.” 585 U.S. at 306, 138 S.Ct. 2206. “The first set of
12a
We’ve then used this precedent to “solidif[y] the line
between short-term tracking of public movements—
akin to what law enforcement could do prior to the
digital age—and prolonged tracking that can reveal
intimate details through habits and patterns.” Leaders
of a Beautiful Struggle v. Balt. Police Dep’t, 2 F.4th 330,
341 (4th Cir. 2021) (en banc) (cleaned up). The latter
“invades the reasonable expectation of privacy that
individuals have in the whole of their movements and
therefore requires a warrant.” Id.
Still, the Supreme Court has recognized that our
existing Fourth Amendment frameworks—like the
third-party doctrine—may be “ill suited to the digital
age,” United States v. Jones, 565 U.S. 400, 417–18, 132
S.Ct. 945, 181 L.Ed.2d 911 (2012) (Sotomayor, J.,
concurring), particularly when applied to cell phones,
which can enable law enforcement to “achieve[ ] near
perfect surveillance,” Carpenter, 585 U.S. at 312, 138
S.Ct. 2206. 5 On top of that, cell phones have become
cases”—including United States v. Knotts, 460 U.S. 276, 103 S.Ct.
1081, 75 L.Ed.2d 55 (1983), and United States v. Jones, 565 U.S. 400,
132 S.Ct. 945, 181 L.Ed.2d 911 (2012)—“address[ed] a person’s
expectation of privacy in his physical location and movements.”
Carpenter, 585 U.S. at 306–07, 138 S.Ct. 2206. “In a second set of
decisions”—including Smith v. Maryland, 442 U.S. 735, 99 S.Ct.
2577, 61 L.Ed.2d 220 (1979)—“the Court [drew] a line between what
a person keeps to himself and what he shares with others,” which is
the guiding principle for the third-party doctrine. Carpenter, 585
U.S. at 307–09, 138 S.Ct. 2206.
5
Even Google—in an amicus brief—argued “that a geofence is
certainly a ‘“search” within the meaning of the Fourth Amendment’
because ‘users have a reasonable expectation of privacy in the
[Location History] information, which the government can use to
13a
“almost ‘a feature of human anatomy’” that individuals
“compulsively carry ... with them all the time.” Id. at 311,
138 S.Ct. 2206 (cleaned up).
So what happens when (as here) there are serious
questions about the scope of a defendant’s consent to a
third-party’s use of his data given the breadth of the
third party’s “detailed, encyclopedic, and effortlessly
compiled” data collection methods? Id. at 309, 138 S.Ct.
2206; see also id. at 315, 138 S.Ct. 2206 (commenting that
exposure of data may not be meaningfully voluntary
when the user doesn’t “‘assume the risk’ of turning over
a comprehensive dossier of his physical movements”
(cleaned up)). Or when (again as here) a “brief snapshot”
of location information, even if it doesn’t capture a
pattern, still “expose[s] highly sensitive information—
think a visit to ‘the psychiatrist, the plastic surgeon, the
abortion clinic, the AIDS treatment center, the strip
club ..., [or] the mosque, synagogue[,] or church’”? Smith,
110 F.4th at 833 (cleaned up); see also Carpenter, 585
U.S. at 311, 138 S.Ct. 2206 (“A cell phone faithfully
follows its owner beyond public thoroughfares and into
private
residences,
doctor’s
offices,
political
headquarters, and other potentially revealing locales.”).
Despite the district court’s best efforts to develop the
record, our understanding of Google’s data collection
policy and its internal geofence warrant process remains
imperfect and incomplete.6 It’s no surprise then that the
retrospectively reconstruct a person’s movements in granular
detail.’” Chatrie, 590 F. Supp. 3d at 907 n.5 (cleaned up).
6
To add more uncertainty, Google intends to change its Location
History policy so that it will no longer be able to respond to geofence
14a
parties vigorously debate—as my colleagues do—the
potentially sweeping implications of any decision.
One camp insists that disallowing geofence warrants
would contravene our precedent, hamstring law
enforcement in investigating crimes, and chill innovation
at any private company that handles a large database of
users. The other camp is just as adamant that granting
blanket approval to these warrants would contravene
our precedent and compromise the privacy interests of
cell phone users.
The balance, ever so delicate, swings from law
enforcement and public safety to liberty and privacy
interests depending on the record facts. Yet despite a
shallow well of information and legal authority and a
litany of unanswered questions as to our decision’s
reach, my colleagues choose to write broadly. At least in
this case, I would opt for restraint and rest on the goodfaith exception to the Fourth Amendment.7
warrants. See Smith, 110 F.4th at 822 n.3.; see also Marlo McGriff,
Updates to Location History and New Controls Coming Soon to
Maps, Google (Dec. 12, 2023), https://blog.google/products/
maps/updates-to-location-history-and-new-controls-coming-soonto-maps/ [https://perma.cc/7ZMS-RHF9].
7
See, e.g., Ashwander v. Tenn. Valley Auth., 297 U.S. 288, 347, 56
S.Ct. 466, 80 L.Ed. 688 (1936) (Brandeis, J., concurring) (“The Court
will not pass upon a constitutional question although properly
presented by the record, if there is also present some other grounds
upon which the case may be disposed of.”); Camreta v. Greene, 563
U.S. 692, 707, 131 S.Ct. 2020, 179 L.Ed.2d 1118 (2011) (“In general,
courts should think hard, and then think hard again, before turning
small cases into large ones.”).
15a
B.
The good-faith exception is reason enough to affirm the
district court without stunting our ability to respond
down the line to Fourth Amendment issues that are
presently “unimaginable.” Chatrie, 590 F. Supp. 3d at
905. Arising out of the exclusionary rule, the exception
broadly queries the deterrent benefits of suppressing an
otherwise constitutionally infirm search. See, e.g., Davis
v. United States, 564 U.S. 229, 236–37, 131 S.Ct. 2419, 180
L.Ed.2d 285 (2011).
Generally, “[t]o trigger the exclusionary rule, police
conduct must be sufficiently deliberate that exclusion
can meaningfully deter it, and sufficiently culpable that
such deterrence is worth the price paid by the justice
system.” Herring v. United States, 555 U.S. 135, 144, 129
S.Ct. 695, 172 L.Ed.2d 496 (2009). And “‘[u]sually, ‘a
warrant issued by a magistrate ... suffices to establish’”
that a law enforcement officer has “acted in good faith in
conducting the search.” Doyle, 650 F.3d at 467 (quoting
Leon, 468 U.S. at 922, 104 S.Ct. 3405).
To better measure any deterrent benefits, courts
consider four circumstances in which good faith won’t
shield even a search made pursuant to a warrant:
(1) If the magistrate or judge in issuing a warrant
was misled by information in an affidavit that the
affiant knew was false or would have known was
false except for his reckless disregard of the
truth; (2) if the issuing magistrate wholly
abandoned his judicial role ... ; (3) if the affidavit
supporting the warrant is so lacking in indicia of
16a
probable cause as to render official belief in its
existence entirely unreasonable; and (4) if under
the circumstances of the case the warrant is so
facially deficient—i.e., in failing to particularize
the place to be searched or the things to be
seized—that the executing officers cannot
reasonably presume it to be valid.
Id. at 467 (cleaned up). None defeat good faith here.
As to the first, Hylton’s occasional sloppiness aside,
there’s no evidence that Hylton gave false information
to the magistrate when seeking the geofence warrant.
And I agree with the government that Chatrie expressly
disclaimed any challenge under Franks v. Delaware, 438
U.S. 154, 98 S.Ct. 2674, 57 L.Ed.2d 667 (1978), that
Detective Hylton “intentionally or recklessly omitted
material information from the affidavit.” See Appellee’s
Br. at 50 (quoting United States v. Pulley, 987 F.3d 370,
376 (4th Cir. 2021)); see also Appellant’s Br. at 11 n.2.
Nor is there evidence that the magistrate didn’t review
the warrant application and Hylton’s affidavit before
issuing the warrant, or that the magistrate at any time
“overstepped his ... judicial responsibilities and
compromised his judicial neutrality.” Chatrie, 590 F.
Supp. 3d at 938 (cleaned up). Chatrie’s citation to Lo-Ji
Sales, Inc. v. New York, 442 U.S. 319, 99 S.Ct. 2319, 60
L.Ed.2d 920 (1979), in which the magistrate became “a
member, if not the leader, of the search party which was
essentially a police operation,” id. at 327, 99 S.Ct. 2319,
is a far cry from the magistrate’s performance here.
At best, Chatrie has “perhaps[ ] shown that [the
magistrate] should have considered the implications of
17a
the [w]arrant more carefully.” Chatrie, 590 F. Supp. 3d
at 938. But our standard for good faith is not so exacting.
The magistrate remained a neutral authority who
reviewed a warrant application describing a novel
investigative tool with a “dearth of court precedent to
follow.” Smith, 110 F.4th at 840.8
Chatrie’s fight isn’t really with the police or the
magistrate. Rather than allege any malfeasance by
either, Chatrie repackages his attack on the warrant’s
probable cause and particularity to suggest that both
acted in bad faith. See, e.g., Appellant’s Br. at 29–30, 32–
33, 38–39. He argues that the warrant was “‘completely
devoid’ of probable cause,” id. at 23, and so “profoundly
lacking in particularity,” id. at 34, as to render it a
“despised” (and illegal) general warrant, id. at 35.
A few points bear repeating. Hylton reviewed
surveillance footage showing that the robber used a cell
phone, so he knew that a geofence could reveal both the
robber’s identity and any potential co-conspirators. The
detective also limited the warrant geographically and
temporally. Hylton, of course, could have further limited
the warrant to a smaller radius around the Bank or a
closer time to the robbery. But given the “dearth of ...
precedent to follow,” Smith, 110 F.4th at 840, nothing
required or cautioned him to do so.
8
Despite holding that geofence warrants are categorically
unconstitutional general warrants, our sister circuit declined to
suppress the evidence under the good-faith exception. Smith, 110
F.4th at 840.
18a
Without any directly governing case law, Hylton
understandably relied on the previous guidance he had
been given, which is, as my colleague explains, “what we
expect reasonable officers to do when faced with such
uncertainty.” Opinion of HEYTENS, J., at 87
(concurring). Magistrates and prosecutors had approved
three of Hylton’s “mostly similar” prior warrants—“all
but one [of which] incorporated a roughly 150-meter
radius.” Chatrie, 590 F. Supp. 3d at 938. As the district
court found, “[e]ven accounting for his miscues, in light
of the complexities of this case, Det[ective] Hylton’s
prior acquisition of three similar warrants, and his
consultation with [g]overnment attorneys before
obtaining those warrants, the [c]ourt cannot say that
[his] reliance on the instant warrant was objectively
unreasonable.” Id. (emphasis added).
Chatrie insists that even a warrant “cloaked” in new
technology must still be supported by probable cause
and be sufficiently particularized as to the places to be
searched and things to be seized. Appellant’s Br. at 24. I
agree with him. But Detective Hylton limited the places
to be searched—both by geography and time—as well as
the location information to be seized—to those cell phone
users within the parameters of the geofence warrant.
To the extent that Chatrie complains that law
enforcement didn’t know his identity in seeking the
warrant (or until well into Google’s three-step process),
I’m not persuaded that carries the day, especially when
assessing good faith. For many warrants, after all, the
point is to identify a suspect, which is why the warrant
requirement focuses on the places to be searched and
things to be seized. See Zurcher v. Stanford Daily, 436
19a
U.S. 547, 555, 98 S.Ct. 1970, 56 L.Ed.2d 525 (1978)
(“Search warrants are not directed at persons; they
authorize the search of ‘places’ and the seizure of
‘things,’ and as a constitutional matter they need not
even name the person from whom the things will be
seized.” (cleaned up)).
Take Zurcher v. Stanford Daily, 436 U.S. 547, 98 S.Ct.
1970, 56 L.Ed.2d 525, in which law enforcement executed
a search warrant of the student newspaper’s offices to
seize “negatives, film, and pictures showing the events
and occurrences at the [Stanford University Hospital]
on the evening” that demonstrators allegedly assaulted
police officers. Id. at 551, 98 S.Ct. 1970. Law enforcement
secured the warrant “on a finding of ‘just, probable and
reasonable cause for believing that’” the things seized—
negatives, photographs, and films—would reveal
“evidence material and relevant to the identity of the
perpetrators.” Id. And the warrant was issued even
though the affidavit “contained no allegation or
indication that members of the Daily staff were in any
way involved in unlawful acts at the hospital.” Id.
No doubt, the initial search here of over 500 million cell
phone users is—to put it mildly—broader than the
search of a handful of college students, but both
warrants were issued to help identify the suspect of the
crime. And in this case, law enforcement narrowed down
the list of potential perpetrators at each step of the
process from millions to dozens to a few based on the
other relevant evidence. That rings in probable cause
sufficient for me to find good faith.
20a
Geofence warrants may differ from traditional warrants,
working in reverse by specifying the time and place of a
crime rather than the identity of the perpetrator, but
that doesn’t automatically render them “facially
deficient,” Doyle, 650 F.3d at 467 (cleaned up). Indeed,
most Internet or mass database searches would be cut
from the same cloth.
All this is to say that it’s not clear what conduct
suppression of the evidence would “meaningfully deter”
here. Herring, 555 U.S. at 144, 129 S.Ct. 695; accord
Chatrie, 590 F. Supp. 3d at 938. Whatever the warrant’s
shortcomings, I agree with the district court that the
warrant wasn’t “so lacking in indicia of probable cause”
as to justify suppressing it here. Chatrie, 590 F. Supp. 3d
at 937 (cleaned up).
III.
When confronted with another opaque and
“transformative” piece of technology, the Supreme
Court recently reminded us that
[t]his challenging new context counsels caution on
our part. As Justice Frankfurter advised 80 years
ago in considering the application of established
legal rules to the “totally new problems” raised
by the airplane and radio, we should take care not
to “embarrass the future.”
TikTok Inc. v. Garland, ––– U.S. ––––, 145 S. Ct. 57, 62,
220 L.Ed.2d 319 (2025) (per curiam) (cleaned up).
My colleagues have done their level best to cut through
the Fourth Amendment fog in this case. In contrast,
21a
some may say that I’ve done nothing more today than
kick the geofence warrant can down the road. Others
may complain that I’ve offered no guidance to law
enforcement and magistrates about the reach of the
Fourth Amendment in the digital age, or worse still, that
I’ve resorted to “judicial abdication,” opinion of WYNN,
J., at 35 (concurring).
But what guidance have my colleagues given today?
Instead of a Fourth Amendment compass, we’ve gifted
law enforcement (and the public) a labyrinth of—by my
count, nine—advisory opinions, many pointing in
different directions.9 See, e.g., Riley v. California, 573
U.S. 373, 398, 134 S.Ct. 2473, 189 L.Ed.2d 430 (2014)
(expressing a “preference” for “provid[ing] clear
guidance to law enforcement” under the Fourth
Amendment); Felix Frankfurter, A Note on Advisory
Opinions, 37 Harv. L. Rev. 1002, 1008 (1942) (“It must
be remembered that advisory opinions are not merely
advisory opinions. They are ghosts that slay.”). I don’t
see the utility in that, as it assumes (wrongly) that we
must give a full answer now.
In short, there are times to make sweeping
constitutional
pronouncements
(with
attendant
9
Even the Fifth Circuit’s opinion, though issued in one voice, has
left legal scholars concerned about its fidelity to the Supreme
Court’s Fourth Amendment precedent, and its implications for all
manner of law enforcement investigative tools. See, e.g., Orin S.
Kerr, The Fifth Circuit Shuts Down Geofence Warrants—And
Maybe a Lot More, The Volokh Conspiracy (August 13, 2024),
https://reason.com/volokh/2024/08/13/fifth-circuit-shuts-down-geof
ence-warrants-and-maybe-a-lot-more/
[https://perma.cc/3G5VWE7F].
22a
consequences) and times to wait. Humility in the face of
the unknown—whether it be the legal ramifications or
practical consequences of our decision, or Google’s own
changing policies—“counsels caution.” TikTok, Inc., 145
S. Ct. at 62.
***
A brief coda. I expect law enforcement to exercise good
faith in using powerful, revolutionary technologies to
investigate crimes, and, indeed, that their first instinct
will be to use and not abuse the information this
technology reveals. And I echo the district court’s
warning that “[d]espite ... finding good faith here, ... this
exception may not carry the day in the future.” Chatrie,
590 F. Supp. 3d at 941.
By my measure, today “our judicial obligation” can “be
captured by a much older rule, familiar to every doctor
of medicine: ‘First, do no harm.’” Denver Area Educ.
Telecomms. Consortium, Inc. v. F.C.C., 518 U.S. 727,
778, 116 S.Ct. 2374, 135 L.Ed.2d 888 (1996) (Souter, J.,
concurring).
23a
WILKINSON, Circuit Judge, with whom NIEMEYER,
KING, AGEE, and RICHARDSON, Circuit Judges,
join, concurring:
With due regard for my fine colleagues, there was no
search here. And even if one were to assume there was
a search, there are many good reasons why courts should
respectfully reject the assault on geofence warrants
mounted by appellant, several of my colleagues, see
opinion of WYNN, J. (concurring), and the Fifth Circuit
Court of Appeals, see United States v. Smith, 110 F.4th
817 (5th Cir. 2024).
I.
There was no search because this case involved a
straightforward application of Smith, 442 U.S. 735, 99
S.Ct. 2577, 61 L.Ed.2d 220 (1979), and Miller, 425 U.S.
435, 96 S.Ct. 1619, 48 L.Ed.2d 71 (1976). Just like in those
cases, Chatrie volunteered incriminating information
about himself to a third party. His expectation of privacy
was comparatively small. Miller, for instance, involved
months of financial transaction history, which
undeniably exposes many intimacies of one’s life. If that
request for bank records was permissible, surely this
request for a two-hour snapshot of one’s public
movements, which hardly reveals one’s habits, is okay.
There are many good reasons why the Supreme Court
did not discard the third-party doctrine for all location
data requests. Of course the concern for privacy in all of
its dimensions was central to the Framers’
contemplation. But the Fourth Amendment, to state the
obvious, calls also for a balance between individual
24a
privacy and public safety. Favoring one over the other is
at odds with the textual “touchstone” of the
Amendment, which is reasonableness. See Maryland v.
King, 569 U.S. 435, 448, 133 S.Ct. 1958, 186 L.Ed.2d 1
(2013). Respecting Fourth Amendment balance means
protecting “that degree of privacy against government
that existed when the Fourth Amendment was
adopted.” Carpenter v. United States, 585 U.S. 296, 305,
138 S.Ct. 2206, 201 L.Ed.2d 507 (2018). Not less, of
course. But also not more.
So yes, the Bill of Rights stands vigilant guard against
the abuses of the state. The Fourth Amendment is itself
a prime illustration of its function. Yet privacy is also
threatened by, say, a theft of personal items. And
privacy is in part a peace of mind. The prospect of
criminal malefactors intruding on that peace can only
mean our privacy has been compromised. That the
transgression is attributable to private actors does not
mean it cannot be part of the calculus of reasonableness
which, again, is our Fourth Amendment touchstone.
Seen in this light, privacy is not invariably in an
adversarial relationship with the state, but something
the state can take measured steps to protect and
provide.
II.
Even if there was a search, there is no room for
emergent judicial hostility toward this new
investigative tool. Disabling the government from using
geofence location data would spurn the basic Fourth
Amendment balance and undermine legitimate law
enforcement in at least three basic ways.
25a
One, this restraint on investigative tools would frustrate
law enforcement’s ability to keep pace with tech-savvy
criminals. Lawless actors of all kinds are growing more
sophisticated and leveraging new technologies to
commit crimes and evade detection. Transnational
criminal organizations rely on digital currencies and
encrypted communications to conceal their violence and
fraud. 2023 WHITE HOUSE STRATEGY TO
COMBAT TRANSNATIONAL ORGANIZED CRIME
3–4, 21 (2023). Terrorists likewise deploy emerging
technologies like encryption, biotechnology, and
artificial intelligence. Ian Moss, U.S. Dep’t of State,
Opening Remarks on Addressing Emerging Technology
in the Realm of Racially or Ethnically Motivated
Violent Extremism (Feb. 14, 2024). Even small-time
pimps encrypt their devices to block lawful access to
their databases of sex-trafficking victims. See Lawful
Access, Office of Legal Policy, U.S. Dep’t of Justice
(Nov. 18, 2022). Examples abound. In this age of
innovation, those who would break the law spare no
expense to employ the latest and greatest technological
tools.
All the while, under appellant’s view, local, state, and
federal officers would lose the tools they need to protect
the public from the modern-day criminal. More cold
cases would go unsolved. Think of a murder where the
culprit leaves behind his encrypted phone and nothing
else. No fingerprints, no witnesses, no murder weapon.
But because the killer allowed Google to track his
location, a geofence warrant can crack the case. See
Damien Christopher & Nick Penzenstadler, Cold Cases
Cracked by Cellphones: How Police Are Using Geofence
Warrants to Solve Crimes, USA TODAY (Sept. 8, 2022).
26a
Taking this tool of last resort out of law enforcement’s
hands would leave these case files collecting dust. The
Fourth Amendment does not require allowing criminals
to take advantage of cutting-edge technologies while
preventing the government from doing the same.
Technology enables the lawbreaker. Courts disable the
government. This imbalance will only grow with time.
Two, law enforcement under appellant’s view would be
robbed of valuable channels of communication with the
private sector. This case is a good example of those
channels at work. Chatrie, like one-third of Google users,
signed up for a program that shared his location data
with Google. In return he got a “virtual journal of his
past travels” and “real-time traffic updates.” United
States v. Chatrie, 107 F.4th 319, 322 (4th Cir. 2024), panel
opinion vacated by order of the en banc court (Nov. 1,
2024). And because he brought his phone to the robbery,
the government was able to place Chatrie at the crime
scene with Google’s help.
Chatrie would shut down this kind of sensible publicprivate cooperation. Doing so would override the
equilibrium between user privacy and public safety that
has emerged organically, without judicial intervention,
from an ecosystem of customers, companies, and law
enforcement. Critics seem to presuppose that private
companies such as Google are naturally disposed to
compromise the privacy of their users. Quite the
contrary. Google has every incentive to protect the
privacy of those who utilize its services. Not to do so
risks damaging its business.
27a
The procedures used by Google here prove the point. In
responding to the government’s location data request,
Google insisted on a rigorous “three-step process” to
protect user privacy. Chatrie, 107 F.4th at 324. It kept
all data anonymized until officers were able to zero in on
a small group of suspects. Only then did Google disclose
the identities of Chatrie and two others. Far from a
“digital dragnet,” the process used here reflected the
reasonable balance between privacy and safety that the
Fourth Amendment envisioned. By urging us to rule
broadly that geofence warrants are impermissible,
Chatrie would unleash a fear of legal liability that would
chill data sharing between public and private sectors and
foreclose fruitful communication over the respective
values of personal privacy and effective law
enforcement.
Three, some of my colleagues go down a dangerous road
by casting the use of geofence data as some new
monster. True, the technology is new, but the technique
is a familiar one. In fact the technique is not too different
from the traditional winnowing methods that criminal
investigators have always used. Investigations often
start out broad. Culprits are not always known, crime
scenes may be crowded, and detectives have to start
somewhere. They canvass the surroundings, review
security footage, and pick out and rule out persons of
interest. Analysis of geofence data follows this same
narrowing progression. So too do keyword searches and
tower dumps. Will courts put a stop to those too? See
Orin Kerr, The Fifth Circuit Shuts Down Geofence
Warrants—And Maybe a Lot More, LAWFARE (Aug.
14, 2024). Will courts seek to disable law enforcement in
cases where there are no eyewitnesses and few forensic
28a
clues? If so, they are far ahead of the Supreme Court in
Carpenter, which ruled on seven days’ worth of location
data, not the snapshot before us now.
III.
There is a further difficulty with categorically
invalidating geofence warrants, namely that of
extending the exclusionary rule with no regard to its
costs. In Hudson v. Michigan, 547 U.S. 586, 126 S.Ct.
2159, 165 L.Ed.2d 56 (2006), the Supreme Court
cautioned against the rule’s “indiscriminate application”
and reiterated that it should apply only when the
“deterrence benefits” outweigh the “social costs.” Id. at
591, 126 S.Ct. 2159. The social costs here are significant.
As we have explained, geofence location data is often the
only way to identify and convict perpetrators like
Chatrie. Excluding this evidence from trial gives these
criminals, in the words of the Supreme Court, “a get-outof-jail-free card.” Id. at 595, 126 S.Ct. 2159. A reflexive
expansion of the exclusionary rule ignores the primary
allegiance of courts to probative evidence and neglects
the Supreme Court’s clear instructions in Hudson.
The creation of remedies involves the weighing of costs
and benefits, which often falls within the domain of
legislators. Indeed, legislatures routinely enact laws
balancing the competing considerations of personal
privacy and public safety. For instance, the Electronic
Communications Privacy Act, Pub. L. No. 99–508, 100
Stat. 1848 (1986), authorizes the government to collect
people’s communications and digital data for law
enforcement purposes. But the law offers a range of
procedural
safeguards—anything
from
an
29a
administrative subpoena to a court-issued warrant
based on probable cause—and remedies depending on
the nature of the data. This type of compromise is a
classic legislative task. Applying the exclusionary rule
categorically to geofence warrants preempts legislative
input in an area whose real impact upon the body politic
would seem to invite some measure of popular
participation.
IV.
As we contemplate the future, Fourth Amendment
interpretation leads to twin risks. One is the risk that
privacy will succumb to the evermore invasive
technological capabilities at the hands of an evermore
intrusive state. The other risk, which is just as real, is
that of privileging those who break the law over those
who would enforce it. Either future portends stark
consequences for society where individual dignity
cannot in the end be divorced from an intuitive sense of
personal safety.
The facts of this case are illustrative. Chatrie terrorized
the employees and patrons of the Call Federal Credit
Union in Midlothian, Virginia. He walked into the bank
armed with a handgun, told the teller that he had
accomplices outside and that he was holding her family
hostage, and threatened to “hurt[ ] everyone in sight” if
she called the cops. United States v. Chatrie, 590 F.
Supp. 3d 901, 905–06 (E.D. Va. 2022). Brandishing his
gun, he forced everyone to the ground and ordered the
manager to empty the safe. Chatrie was able to escape
with $195,000. Because he was not apprehended at the
scene, he eluded law enforcement for months. Officers
30a
were out of traditional leads. Only the geofence warrant
eventually allowed police to track Chatrie down and
restore a sense of resolution to the community. Without
geofence location data, crimes even more serious than
this one will escape detection.
The sheer breadth of appellant’s position is disquieting.
Those who support it seek a broad judicial declaration
that geofence warrants would be unconstitutional in all
their forms, no matter how specific and particularized.
The geofence warrant here was closely confined to a
particular time, place, and incident. There can be abuses
to be sure, but courts can review the temporal and
spatial character of these warrants as we would any
Fourth Amendment claim. To strike the warrant down
here comes pretty nearly to invalidating it everywhere.
No matter says appellant. All such warrants are on the
chopping block.
Crime invades privacy. Crime limits freedom and
narrows space. The fact that the Fourth Amendment
exists to check the undeniable excesses of the modern
state does nothing to diminish the fact that crime
imperils the very values the Fourth Amendment exists
to protect. The Framers resolved this dilemma by
making reasonableness the Amendment’s touchstone. It
is dispiriting that some would proceed with nary a
thought given to that two-sided balance which
reasonableness above all denotes. It will never do to see
the future with but a single eye.
31a
NIEMEYER, Circuit Judge, concurring:
I am pleased to join the opinions of Judge Wilkinson and
Judge Richardson in full. Today’s Fourth Amendment
caselaw often starts with a pre-Internet analogy. See
Carpenter v. United States, 585 U.S. 296, 306, 138 S.Ct.
2206, 201 L.Ed.2d 507 (2018). I write separately because
I believe that a commonsense analogy dictates the same
result reached by the opinions of Judge Wilkinson and
Judge Richardson.
To begin, the Fourth Amendment protects the people
“in their persons, houses, papers, and effects” against
unreasonable searches. U.S. Const. amend. IV. It has
also been construed to extend beyond those textual
objects to protect certain expectations of privacy. See
Katz v. United States, 389 U.S. 347, 353, 88 S.Ct. 507, 19
L.Ed.2d 576 (1967); id. at 361, 88 S.Ct. 507 (Harlan, J.,
concurring). And recently, the Supreme Court held in
Carpenter that the Fourth Amendment protects a
person’s expectation of privacy in “the whole of his
physical movements.” 585 U.S. at 313, 138 S.Ct. 2206.
Thus, when law enforcement, without a warrant,
accesses a person’s continuously collected and
automatically generated cell-site location information, it
violates that expectation of privacy. See id. at 315–16,
138 S.Ct. 2206. But Carpenter left in place many existing
limits on the scope of the Fourth Amendment. Apart
from protecting the unique data-collection system at
issue there, the Carpenter Court explained that it was
not “disturb[ing] the application” of the third-party
doctrine “or call[ing] into question conventional
surveillance techniques and tools, such as security
cameras. Nor [did it] address other business records that
32a
might incidentally reveal location information.” Id. at
316, 138 S.Ct. 2206.
One of the “conventional surveillance techniques” that
Carpenter left untouched is law enforcement’s practice
of collecting and following “markers,” or clues,
voluntarily left behind and abandoned by a person at the
scene of a crime or in connection with the crime. These
markers can reveal who committed the crime, and, when
the crime was committed in a public place or in the place
of a third person, they may be collected by law
enforcement without a warrant. Thus, law enforcement
is entitled to retrieve boot prints, tire tracks, shell
casings, a scarf or a cap, and items left with fingerprints
or DNA on them. Similarly, they can retrieve thirdperson records of a suspect’s presence, such as pictures
and videos taken routinely at the scene, records of tolls
paid, or records of credit card transactions. Indeed, such
third-party records might include a note left with a teller
during a bank robbery. Collecting markers such as these
from public places or third persons is the stuff of law
enforcement, enabling it to solve crimes and prosecute
suspects, and the person who left them behind is not
“searched” in his person and effects, in violation of the
Fourth Amendment.
Of course, if a person were careful not to leave
footprints, fingerprints, shell casings, or other markers
behind, law enforcement would have to turn to other
techniques and strategies to advance its investigation.
But when such markers are left behind, law enforcement
should not be denied the benefit of the person’s
carelessness when solving a crime. And Carpenter says
nothing to the contrary. What Carpenter does say is that
33a
law enforcement needs to obtain a warrant before it
utilizes digital technology to track a citizen’s long-term
movements — “the whole of his physical movements” —
at least when that person is, in effect, compelled to leave
behind a digital footprint wherever he goes. 585 U.S. at
313, 315, 138 S.Ct. 2206; see also United States v. Jones,
565 U.S. 400, 430, 132 S.Ct. 945, 181 L.Ed.2d 911 (2012)
(Alito, J., concurring in the judgment); Leaders of a
Beautiful Struggle v. Balt. Police Dep’t, 2 F.4th 330, 341
(4th Cir. 2021) (en banc). But those features are not
present here, and, as this case is otherwise well-removed
from the text of the Fourth Amendment, I would hold
that law enforcement did not conduct a search.
This case relates to law enforcement’s effort to collect
markers from third persons voluntarily left behind by a
person during the commission of a crime. In this case,
the person left behind electronic location data that he
voluntarily transmitted from the scene of the crime by
his cell phone. Law enforcement did not collect the data
from the person or the person’s cell phone, which would
require a warrant, see Riley v. California, 573 U.S. 373,
401, 134 S.Ct. 2473, 189 L.Ed.2d 430 (2014), but from a
third person who received the person’s voluntarily
transmitted data and stored them in a data bank, see
Smith v. Maryland, 442 U.S. 735, 743–44, 99 S.Ct. 2577,
61 L.Ed.2d 220 (1979); United States v. Miller, 425 U.S.
435, 443, 96 S.Ct. 1619, 48 L.Ed.2d 71 (1976). In this
sense, the data, when limited to the time and place of the
crime, were no different than any other marker left
behind by a perpetrator.
What might distinguish such electronic data from other
markers is the scope of the data collection. Here, the
34a
data were retained by the third person in a large data
bank — Google’s Sensorvault — which includes
information unrelated to the time and place of the crime.
The broad scope of that data bank could raise privacy
concerns for those whose data were stored there,
including the suspect’s data that did not constitute a
marker from the crime scene. But law enforcement
accessed only two hours’ worth of location data, which is
far from “the whole of [anyone’s] physical movements.”
Carpenter, 585 U.S. at 313, 138 S.Ct. 2206. And law
enforcement relied on procedures designed to isolate the
data constituting markers left behind at the crime scene
from other, unrelated data, which helped mitigate any
privacy concerns.
The geofence warrant issued in this case initially
required Google to produce data transmitted by cell
phones only (1) from the scene of the crime and (2)
during the time when the crime was committed. They
were thus potential crime markers, which helped law
enforcement solve the crime and were not materially
distinct from the fingerprints or shell casings left behind
by a prior era’s less-than-careful perpetrators.
At bottom, this case is a good example of law
enforcement properly balancing its need to solve and
prosecute crimes with citizens’ privacy concerns under
the Fourth Amendment. Neither the suspect nor any
other person whose data was stored in the data bank
could legitimately claim, in view of the procedures
followed, that his rights were violated. Judge
Richardson’s opinion neatly, systematically, and
accurately sets forth the legal principles supporting this
conclusion, and Judge Wilkinson’s opinion elegantly
35a
articulates the public policies that this conclusion
promotes.
In addition, I also concur in the judgment of the court
holding that, in any event, law enforcement’s collection
of the data from Google was protected because law
enforcement relied in good faith on a warrant issued by
a detached and neutral judicial officer. See United States
v. Leon, 468 U.S. 897, 922–23, 104 S.Ct. 3405, 82 L.Ed.2d
677 (1984).
36a
KING, Circuit Judge, concurring:
I am pleased to join in the fine concurring opinions of
Judge Wilkinson and Judge Richardson. In addition, I
agree that the officers acted in good faith, and I
therefore also support the affirmance of the district
court’s judgment on that basis.
37a
WYNN, Circuit Judge, with whom Judges THACKER,
HARRIS, BENJAMIN, and BERNER join, and with
whom Judge GREGORY joins except as to footnote 1,
concurring in the judgment:
The surveillance technologies at issue in this case—the
very same ones that seem to thrill my colleagues who
join Judge Wilkinson’s separate opinion—would have
been unimaginable to the Founders. Yet, in Carpenter v.
United States, 585 U.S. 296, 138 S.Ct. 2206, 201 L.Ed.2d
507 (2018), our Supreme Court rightly recognized that
the principles enshrined in the Fourth Amendment do
not wither in the face of advancing technologies. Rather,
they must be vigorously protected from ever-expanding
methods of government intrusion.
The Court in Carpenter reaffirmed a fundamental truth:
until, and unless, the Constitution is amended, it is the
duty of the judiciary to defend constitutional rights
against encroachments that the Framers could not have
foreseen but surely would have found intolerable.
Thus, “when a Fourth Amendment case presents a novel
question of law whose resolution is necessary to guide
future action by law enforcement officers and
magistrates, there is sufficient reason for [a court] to
decide the violation issue before turning to the goodfaith question.” United States v. Bosyk, 933 F.3d 319, 332
n.10 (4th Cir. 2019) (quoting Illinois v. Gates, 462 U.S.
213, 264, 103 S.Ct. 2317, 76 L.Ed.2d 527 (1983) (White, J.,
concurring)); see United States v. Leon, 468 U.S. 897, 925,
104 S.Ct. 3405, 82 L.Ed.2d 677 (1984). “As demonstrated
by the divergent decisions of district courts”—and here,
38a
of circuit courts—“this is one such case.” Bosyk, 933 F.3d
at 332 n.10.
The constitutional question in this case has been fully
briefed, argued and exhaustively debated—not only by
the parties but by amici and members of this Court. And
it is unclear what future case could better tee up the
issue. Judicial modesty does not demand judicial
abdication.
Yet, by declining to reach the merits in this matter, this
Court squanders a critical opportunity to clarify the
Fourth Amendment’s application to emerging
surveillance technologies. Instead, we take shelter in the
judge-made doctrine of “good faith,” leaving both courts
and citizens to grope in the dark as to the limits of
governmental power in the digital age. The result?
Individuals subject to sweeping, sophisticated
surveillance with little or no judicial oversight—an
outcome wholly at odds with our constitutional design.
I therefore write separately to explain why, in obtaining
Google Location History data traceable to Okello
Chatrie, the police conducted a Fourth Amendment
search.1
I.
The Fourth Amendment promises “secur[ity] ... against
unreasonable searches and seizures.” U.S. Const.
1
Although I believe that this case involved a Fourth Amendment
search—and that we should say so—I acknowledge that the
conditions for application of the good-faith exception to the
exclusionary rule are met here.
39a
amend. IV. Surveillance technologies, though also
deployed in the name of security, pose a dynamic and
resilient threat to that right. Technology continually
advances; consequently, maintaining the balance
between individual privacy and public safety requires
vigilance. Recognizing this, the Supreme Court has
allowed Fourth Amendment jurisprudence to evolve
alongside technology. I begin by surveying that
evolution, with particular attention to its latest chapter:
the Court’s decision in Carpenter.
A.
Early Supreme Court decisions made clear that a
government agent’s physical trespass into a private
space is a search, and thus requires a warrant. But as the
Government’s capacity to surveil at a distance
expanded, so did the Fourth Amendment’s protections.
See Carpenter, 585 U.S. at 304, 138 S.Ct. 2206. The
modern rule—adapted from Justice Harlan’s concurring
opinion in Katz v. United States, 389 U.S. 347, 88 S.Ct.
507, 19 L.Ed.2d 576 (1967), and reaffirmed many times
since—is that “[w]hen an individual seeks to preserve
something as private, and his expectation of privacy is
one that society is prepared to recognize as reasonable,
... official intrusion into that private sphere generally
qualifies as a search and requires a warrant supported
by probable cause.” Carpenter, 585 U.S. at 304, 138 S.Ct.
2206 (internal quotation marks omitted).
In the 1970s and 1980s—before the internet age—the
Supreme Court placed two key limitations on Katz’s
expansion of recognized Fourth Amendment
protections: the third-party and public-surveillance
40a
doctrines. See id. at 306–09, 138 S.Ct. 2206.
Understanding those limitations is essential to
understanding the Court’s later decision in Carpenter.
First, the third-party doctrine stems from decisions
issued over 45 years ago: Smith v. Maryland, 442 U.S.
735, 99 S.Ct. 2577, 61 L.Ed.2d 220 (1979), and United
States v. Miller, 425 U.S. 435, 96 S.Ct. 1619, 48 L.Ed.2d
71 (1976).
In Smith, police used a pen-register device to collect
phone numbers the suspect dialed on his home phone.
Smith, 442 U.S. at 737, 99 S.Ct. 2577. And in Miller,
police accessed the suspect’s bank records, such as
checks and deposit slips. Miller, 425 U.S. at 437–38, 96
S.Ct. 1619. The Supreme Court held that the suspects
had no reasonable expectation of privacy in those
relatively unrevealing records, which the suspects had
voluntarily exposed to third parties in the ordinary
course of business. See Smith, 442 U.S. at 737, 741–42, 99
S.Ct. 2577; Miller, 425 U.S. at 440–43, 96 S.Ct. 1619;
Carpenter, 585 U.S. at 308–09, 138 S.Ct. 2206 (discussing
Smith and Miller).
Second, the public-surveillance doctrine emerges from
decisions issued over 40 years ago, and centers on
differing expectations of privacy in public versus private
spaces.
In United States v. Knotts, 460 U.S. 276, 103 S.Ct. 1081,
75 L.Ed.2d 55 (1983), the Court held that police did not
conduct a Fourth Amendment search when they used a
“beeper”—that is, “a radio transmitter” that “emits
periodic signals that can be picked up by a radio
41a
receiver”—to keep a vehicle in view during a single
drive “on public thoroughfares.” Id. at 277, 281, 103 S.Ct.
1081. The Court reasoned that police could have tracked
the vehicle’s movements without the beeper—by
physically following it—so the suspect had no reasonable
expectation of privacy in those movements. Id. at 281–
82, 285, 103 S.Ct. 1081.
Knotts “was careful to distinguish between the
rudimentary tracking facilitated by the beeper and more
sweeping modes of surveillance.” Carpenter, 585 U.S. at
306, 138 S.Ct. 2206. The Court stressed that the beeper
merely “augment[ed]” the officers’ own “sensory
faculties.” Knotts, 460 U.S. at 282, 103 S.Ct. 1081. And it
cautioned that, should “twenty-four hour surveillance of
any citizen” become “possible,” “different constitutional
principles may be applicable.” Carpenter, 585 U.S. at
306–07, 138 S.Ct. 2206 (quoting Knotts, 460 U.S. at 283–
84, 103 S.Ct. 1081).
The Court distinguished Knotts in United States v.
Karo, 468 U.S. 705, 104 S.Ct. 3296, 82 L.Ed.2d 530 (1984),
which held that police conducted a Fourth Amendment
search when they used a beeper to track a container as
it moved between commercial lockers and private
residences. Id. at 708–10, 714–18, 104 S.Ct. 3296. The
Court explained that because “private residences are
places in which the individual normally expects privacy
free of governmental intrusion not authorized by a
warrant, and that expectation is plainly one that society
is prepared to recognize as justifiable,” “[s]earches and
seizures inside a home without a warrant are
presumptively
unreasonable
absent
exigent
circumstances.” Id. at 714–15, 104 S.Ct. 3296. Although
42a
tracking the beeper’s location was “less intrusive than a
full-scale search,” it “reveal[ed] a critical fact about the
interior of the premises”; and unlike the public
movements of the vehicle in Knotts, police “could not
have otherwise obtained [that information] without a
warrant.” Id. at 715, 104 S.Ct. 3296.
In short, Smith, Miller, Knotts, and Karo—all decided
before 1985—recognized that there is no reasonable
expectation of privacy in simple records voluntarily
conveyed to third parties in the ordinary course of
business, or in one’s short-term public movements. But
as new surveillance technologies “enhanced the
Government’s capacity to encroach upon areas normally
guarded from inquisitive eyes,” the Supreme Court
“sought to ‘assure [ ] preservation of that degree of
privacy against government that existed when the
Fourth Amendment was adopted.’” Carpenter, 585 U.S.
at 305, 138 S.Ct. 2206 (quoting Kyllo v. United States,
533 U.S. 27, 34, 121 S.Ct. 2038, 150 L.Ed.2d 94 (2001)).
Three cases illustrate that endeavor.
First, Kyllo v. United States held that police use of a
thermal-imaging device to monitor heat waves
emanating from inside a home was a search. Kyllo, 533
U.S. at 34–35, 121 S.Ct. 2038. The Court explained that
even though the device was operated from a public
street outside the home, it allowed police to “explore
details of the home that would previously have been
unknowable without physical intrusion.” Id. at 40, 121
S.Ct. 2038. “Because any other conclusion would leave
homeowners ‘at the mercy of advancing technology,’”
the Court “determined that the Government—absent a
warrant—could not capitalize on such new sense-
43a
enhancing technology to explore what was happening
within the home.” Carpenter, 585 U.S. at 305, 138 S.Ct.
2206 (quoting Kyllo, 533 U.S. at 35, 121 S.Ct. 2038).
Next, in United States v. Jones, 565 U.S. 400, 132 S.Ct.
945, 181 L.Ed.2d 911 (2012), the Court grappled with
“more sophisticated surveillance of the sort envisioned
in Knotts and found that different principles did indeed
apply.” Carpenter, 585 U.S. at 307, 138 S.Ct. 2206. Jones
held that the police’s installation and use of a GPS
tracking device to monitor the location of a suspect’s
vehicle for 28 days constituted a search. Jones, 565 U.S.
at 403–04, 132 S.Ct. 945. Although Justice Scalia’s
opinion for the five-justice majority rested only on
traditional trespass principles, five other justices
authored or joined concurrences concluding that the
GPS monitoring was a search under Katz’s reasonableexpectation-of-privacy test—even though the intrusion
only captured public movements. See id. at 413–18, 132
S.Ct. 945 (Sotomayor, J., concurring); id. at 418–31, 132
S.Ct. 945 (Alito, J., concurring in the judgment). The
concurring justices noted that, as compared to the onetrip beeper tracking in Knotts, the GPS tracking in
Jones was both longer and more precise. See id. at 415,
132 S.Ct. 945 (Sotomayor, J., concurring); id. at 429–30,
132 S.Ct. 945 (Alito, J., concurring in the judgment).
Specifically, four concurring justices emphasized that
long-term
GPS
tracking
violated
reasonable
expectations of privacy because it enabled police to tail
a suspect for much longer than would have been possible
using traditional investigative methods. See id. at 429,
132 S.Ct. 945 (Alito, J., concurring in the judgment) (“In
the pre-computer age, the greatest protections of
44a
privacy were neither constitutional nor statutory, but
practical.”).
For the fifth concurring justice, Justice Sotomayor, even
short-term GPS tracking violated reasonable
expectations of privacy because it enabled such precise
surveillance. Id. at 415, 132 S.Ct. 945 (Sotomayor, J.,
concurring). She reasoned that GPS technology
“generates a precise, comprehensive record of a person’s
public movements that reflects a wealth of detail about
her familial, political, professional, religious, and sexual
associations.” Id. And because a short GPS search is
cheaper, easier to use, and more concealable than
conventional methods of surveillance, “it evades the
ordinary checks that constrain abusive law enforcement
practices: ‘limited police resources and community
hostility.’” Id. at 416, 132 S.Ct. 945 (quoting Illinois v.
Lidster, 540 U.S. 419, 426, 124 S.Ct. 885, 157 L.Ed.2d 843
(2004)). Moreover, GPS technology permits the
Government to “store” and “efficiently mine” records of
an individual’s movements for “years into the future.”
Id. at 415, 132 S.Ct. 945. For these reasons, even a short
GPS search could chill First Amendment freedoms and
“alter the relationship between citizen and government
in a way that is inimical to democratic society.” Id. at 416,
132 S.Ct. 945 (citation omitted).2
2
“More fundamentally,” Justice Sotomayor argued, “it may be
necessary to reconsider the premise that an individual has no
reasonable expectation of privacy in information voluntarily
disclosed to third parties.” Jones, 565 U.S. at 417, 132 S.Ct. 945. That
“approach is ill suited to the digital age, in which people reveal a
great deal of information about themselves to third parties in the
course of carrying out mundane tasks,” without expecting their
45a
Two years later, the Court held in Riley v. California,
573 U.S. 373, 134 S.Ct. 2473, 189 L.Ed.2d 430 (2014), that
police must obtain a warrant to look through the
contents of an arrestee’s cell phone during an arrest,
even though police may generally conduct brief searches
of an arrestee’s person without a warrant. Id. at 385–86,
134 S.Ct. 2473. The Court recognized that digital storage
compiles personal information of unprecedented volume,
variety, and retrospectivity into a single device (or, in
the Fourth Amendment’s language, “effect”)—and
consequently, that protecting privacy rights in such
effects require a different approach. Id. at 393–97, 134
S.Ct. 2473.
In each of these seminal cases, the Supreme Court
grappled with how to protect constitutional privacy
rights from encroaching technologies. And, in the
majority opinions in most of these cases and in the Jones
concurrences, the Court recognized that then-existing
Fourth Amendment case law was ill-adapted to the
realities of modern technology.
B.
The Court’s growing recognition of the profound impact
of technological advancements on Fourth Amendment
rights was on full display in its 2018 decision in
Carpenter v. United States. While building on all that
came before it, Carpenter marked a “sea change” in
Fourth Amendment jurisprudence as it pertains to “a
person’s digital information.” Matthew Tokson, The
devices “to enable covert surveillance of their movements.” Id. at
417 & n.*, 132 S.Ct. 945.
46a
Aftermath of Carpenter: An Empirical Study of Fourth
Amendment Law, 2018–2021, 135 Harv. L. Rev. 1790,
1799–1800 (2022) (capitalization standardized).
In Carpenter, the Court held that law enforcement’s
request for seven days of the defendant’s historical cellsite location information (“CSLI”) from his wireless
carrier, which produced two days’ worth of data, was a
search. Carpenter, 585 U.S. at 302, 316, 138 S.Ct. 2206.
CSLI records are created when cell phones connect to
nearby cell towers, which, in Carpenter, occurred at the
start and end of the defendant’s incoming and outgoing
calls. Id. at 302, 138 S.Ct. 2206. The cell-site records were
maintained by wireless carriers, which raised the
possibility that the third-party doctrine would apply.
And indeed, below, the Sixth Circuit had “held that [the
defendant] lacked a reasonable expectation of privacy in
the location information collected by the FBI because he
had shared that information with his wireless carriers.”
Id. at 303, 138 S.Ct. 2206; see United States v. Carpenter,
819 F.3d 880, 888 (6th Cir. 2016). In other words, the
Sixth Circuit took a view very similar to that of some of
my colleagues here. See Judge Richardson Concurring
Op., infra, Part II.B.
But the Supreme Court reversed. In so doing, it
acknowledged that the third-party doctrine is an
increasingly tenuous barometer for reasonable privacy
expectations in the digital era. Instead, the Court laid
the foundation for a new, multifactor test to determine
when
government
surveillance
using
digital
technologies constitutes a search.
47a
Carpenter began with the Katz test: the Fourth
Amendment protects against intrusion into the sphere
in which an individual has a reasonable expectation of
privacy. Carpenter, 585 U.S. at 304, 138 S.Ct. 2206. It
then explained that, while “no single rubric” defines
reasonable expectations of privacy, the Court’s analysis
must always be “informed by historical understandings
of what was deemed an unreasonable search when the
Fourth Amendment was adopted.” Id. at 304–05, 138
S.Ct. 2206 (cleaned up). These historical understandings,
according to the Court, have a few “guideposts”: “the
[Fourth] Amendment seeks to secure the privacies of
life against arbitrary power,” “to place obstacles in the
way of a too permeating police surveillance,” and, most
importantly, to “assure preservation of that degree of
privacy against government that existed when the
Fourth Amendment was adopted.” Id. at 305, 138 S.Ct.
2206 (cleaned up). The Court emphasized that in cases
like Kyllo and Riley, it kept those “Founding-era
understandings in mind” when considering “innovations
in surveillance tools.” Id.
Against that background, the Court quickly concluded
that CSLI—“personal location information maintained
by a third party”—“does not fit neatly” into any existing
line of Fourth Amendment jurisprudence. Id. at 306, 138
S.Ct. 2206. The third-party-disclosure and publicsurveillance cases could “inform [the Court’s]
understanding of the privacy interests at stake,” but
neither squarely applied. Id. In fact, the Court expressly
“decline[d] to extend” the third-party doctrine to
CSLI—even though CSLI data is maintained by thirdparty companies—because CSLI records are a
“qualitatively different category” of information from
48a
the phone numbers and bank records at issue in its thirdparty cases. Id. at 309, 138 S.Ct. 2206. “After all,” the
Court observed, “when Smith was decided in 1979, few
could have imagined a society in which a phone goes
wherever its owner goes, conveying to the wireless
carrier not just dialed digits, but a detailed and
comprehensive record of the person’s movements.” Id.
at 309, 138 S.Ct. 2206.
Instead of “mechanically applying the third-party
doctrine,” id. at 314, 138 S.Ct. 2206, Carpenter applied a
new framework rooted in historical understandings of
Fourth Amendment privacy rights but adapted to the
particular surveillance technology at issue. Specifically,
the Court identified four aspects of CSLI surveillance
that made it “qualitatively different” from older
techniques—its comprehensiveness, its capacity for
retrospective tracking, the intimacy of the information it
reveals, and its ease of access for police.3 See id. at 309–
13, 138 S.Ct. 2206. Based on those four considerations,
the Court concluded that police access to CSLI violates
reasonable expectations of privacy. Id. at 313, 138 S.Ct.
2206.
Then, in a separate section of the opinion, the Court
further distinguished Smith and Miller by explaining
that the conveyance of CSLI is also not meaningfully
3
Carpenter’s framework drew on the reasoning of the Jones
concurrences, and particularly Justice Sotomayor’s concurrence. Cf.
Jones, 565 U.S. at 415–16, 132 S.Ct. 945 (Sotomayor, J., concurring)
(observing that “GPS monitoring generates a precise,
comprehensive record” of “intimate information” that can be
“store[d]” and “efficiently mine[d] ... for information years into the
future”).
49a
voluntary. Id. at 313–16, 138 S.Ct. 2206. The opinion’s
concluding paragraph reads, in part: “In light of the
deeply revealing nature of CSLI, its depth, breadth, and
comprehensive reach, and the inescapable and automatic
nature of its collection, the fact that such information is
gathered by a third party does not make it any less
deserving of Fourth Amendment protection.” Id. at 320,
138 S.Ct. 2206.
II.
Carpenter established a multifactor approach to
assessing reasonable expectations of privacy in digital
information.4 An application of the Carpenter factors in
4
Leading scholars agree, though they differ as to which factors are
mandatory or most important. See, e.g., Paul Ohm, The Many
Revolutions of Carpenter, 32 Harv. J.L. & Tech. 357, 363, 369 (2019)
(recognizing that Carpenter created “new, multi-factor test” to
analyze an individual’s reasonable privacy expectation against
intruding technology and “herald[ed] a new mode of Constitutional
analysis”); Susan Freiwald & Stephen W. Smith, The Carpenter
Chronicle: A Near-Perfect Surveillance, 132 Harv. L. Rev. 205, 219
(2018) (multifactor analysis was “clearly central” to the Court’s
holding); Tokson, The Aftermath of Carpenter, supra, at 1830
(describing the “Carpenter factors” and concluding from a survey of
cases that “[a] multifactor Carpenter test has begun to emerge from
the lower court[s]”); Sherwin Nam, Bend and Snap: Adding
Flexibility to the Carpenter Inquiry, 54 Colum. J.L. & Soc. Probs.
131, 132 (2020) (stating that Carpenter “broke new ground in the
constitutional right to privacy in electronic data” and employed a
“five-factor” test); Helen Winters, An (Un)reasonable Expectation
of Privacy? Analysis of the Fourth Amendment When Applied to
Keyword Search Warrants, 107 Minn. L. Rev. 1369, 1381, 1390
(2023) (Carpenter “marked a new period of Fourth Amendment
jurisprudence” and laid out “several factors relevant to its
decision”); Antony Barone Kolenc, “23 and Plea”: Limiting Police
Use of Genealogy Sites After Carpenter v. United States, 122 W.
50a
this case compels the conclusion that Okello Chatrie had
a reasonable expectation of privacy in his Location
History data.5
A.
Carpenter first considered the comprehensiveness of
CSLI data, observing that it “tracks nearly exactly the
movements of [a cell phone’s] owner,” providing “an allencompassing record of the holder’s whereabouts.”
Carpenter, 585 U.S. at 311, 138 S.Ct. 2206. Unlike a
vehicle, “a cell phone—almost a ‘feature of human
anatomy’— ... faithfully follows its owner beyond public
thoroughfares and into private residences, doctor’s
Va. L. Rev. 53, 71–72 (2019) (concluding that Carpenter “alter[ed]
Fourth Amendment law” by recognizing a privacy interest in the
“whole of a person’s physical movements,” and “balanced five
factors” to analyze that interest); Matthew Tokson, The Carpenter
Test as a Transformation of Fourth Amendment Law, 2023 U.
Illinois L. Rev. 507, 517–20 (2023) (outlining a three-factor test);
Allie Schiele, Learning from Leaders: Using Carpenter to Prohibit
Law Enforcement Use of Mass Aerial Surveillance, 91 Geo. Wash.
L. Rev. Arguendo 14, 17–18 (2023) (pointing out “Carpenter’s focus
on five central factors”); Nicole Mo, If Wheels Could Talk: Fourth
Amendment Protections Against Police Access to Automobile
Data, 98 N.Y.U. L. Rev. 2232, 2251 (2023) (recognizing factors);
Luiza M. Leão, A Unified Theory of Knowing Exposure:
Reconciling Katz and Carpenter, 97 N.Y.U. L. Rev. 1669, 1684
(2022) (same); Matthew E. Cavanaugh, Somebody’s Tracking Me:
Applying Use Restrictions to Facial Recognition Tracking, 105
Minn. L. Rev. 2443, 2468 (2021) (same).
5
Police obtained Chatrie’s Location History data when they
obtained Location History data that was traceable to him. Here—
as Judge Berner persuasively explains—that happened at Step 2 of
Google’s three-step process. See Judge Berner Concurring Op.,
infra, Part II.B.i.
51a
offices, political headquarters, and other potentially
revealing locales.” Id. (quoting Riley, 573 U.S. at 385,
134 S.Ct. 2473).
Like CSLI, Location History tracks a smartphone’s
location—only more precisely. CSLI (as described in
Carpenter) places a user within a “wedge-shaped
sector,” id. at 312, 138 S.Ct. 2206, ranging from “a dozen”
to “several hundred” city blocks in size, which can be “up
to 40 times more imprecise” in rural areas, id. at 324, 138
S.Ct. 2206 (Kennedy, J., dissenting). But Location
History can locate a user within meters—and can even
measure elevation, identifying the specific floor in a
building where a person might be. United States v.
Chatrie, 590 F. Supp. 3d 901, 908–09 (E.D. Va. 2022).
Moreover, the CSLI collected in Carpenter was only
recorded when a user placed or received a call—no call,
no data. Carpenter, 585 U.S. at 302, 138 S.Ct. 2206. But
Location History tracks a user’s location automatically,
every two minutes. Chatrie, 590 F. Supp. 3d at 908. In
Carpenter, law enforcement collected only about 101
CSLI data points in a full day. Carpenter, 585 U.S. at
302, 138 S.Ct. 2206. Here, police were able to collect an
average of about 76 Location History data points on each
person surveilled in just two hours. See J.A. 1121
(explaining that “Google produced ... a total of 680 data
points” for “nine accounts” at Step 2). If CSLI as
described in Carpenter enables “near perfect
surveillance,” Carpenter, 585 U.S. at 312, 138 S.Ct. 2206,
so too does Location History.
52a
B.
Carpenter next considered “the retrospective quality of
[CSLI] data,” which (at the time) was “continually
logged for all of the 400 million devices in the United
States” and retained by wireless carriers “for up to five
years.” Carpenter, 585 U.S. at 312, 138 S.Ct. 2206. CSLI
allowed police to “travel back in time” to “reconstruct a
person’s movements,” unlocking “a category of
information otherwise unknowable.” Id. And because
CSLI tracking “runs against everyone,” “police need not
even know in advance whether they want to follow a
particular individual, or when.” Id. “Whoever the
suspect turns out to be, he has effectively been tailed
every moment of every day for five years.” Id.
Location History data raises similar concerns. Google
begins collecting Location History the moment the
feature is enabled and retains it indefinitely, enabling
police to retrospectively tail a suspect with remarkable
precision.6 And like CSLI, police need not identify the
suspect in advance—Location History data is available
for “numerous tens of millions” of Google users. Chatrie,
590 F. Supp. 3d at 907. Of course, a geofence limits the
size and duration of any particular law enforcement
data-grab. But Carpenter’s retrospectivity analysis
emphasized the vast scope of available CSLI data, which
gives police “access to a category of information
otherwise unknowable.” Carpenter, 585 U.S. at 312, 138
S.Ct. 2206 (emphasis added). So too here.
6
This discussion reflects the record in this case, not Google’s
current or future practices.
53a
In fact, Location History permits even broader
surveillance than CSLI. Collecting CSLI data at least
requires police to produce a suspect’s phone number in
order to access a five-year trove of their location data.
But a geofence can uncover the Location History of an
unlimited number of individuals, none of whom were
previously identified or suspected of any wrongdoing.
Indeed, the very point of a geofence is to generate leads
where none exist.7 Consequently, Carpenter’s concerns
about retrospective surveillance apply to Location
History with even greater force.
C.
Carpenter further concluded that “time-stamped
[location] data provides an intimate window into a
person’s life, revealing not only his particular
movements, but through them his ‘familial, political,
professional, religious, and sexual associations.’”
Carpenter, 585 U.S. at 311, 138 S.Ct. 2206 (quoting
Jones, 565 U.S. at 415, 132 S.Ct. 945 (Sotomayor, J.,
7
This feature of geofence warrants makes them uncomfortably
akin to the “reviled” general warrants that the Framers intended
the Fourth Amendment to forbid. Carpenter, 585 U.S. at 303, 138
S.Ct. 2206 (quoting Riley, 573 U.S. at 403, 134 S.Ct. 2473); see
United States v. Smith, 110 F.4th 817, 836–38 (5th Cir. 2024). “The
general warrant specified only an offense ... and left to the discretion
of the executing officials the decision as to which persons should be
arrested and which places should be searched.” Steagald v. United
States, 451 U.S. 204, 220, 101 S.Ct. 1642, 68 L.Ed.2d 38 (1981). As
Judge Berner explains, probable cause may support a tightly
limited geofence warrant. See Judge Berner Concurring Op., infra,
Part II.D. But if accessing Location History is not a search at all,
police would not even need to specify an offense before dipping into
years of personal location data on millions of Americans.
54a
concurring)). Such “location records,” the Court
recognized, “hold for many Americans the privacies of
life.” Id. (quoting Riley, 573 U.S. at 403, 134 S.Ct. 2473).
The same is true of Location History. The two hours of
geographically unbounded data requested by police at
Step 2 illustrate that “the potential intrusiveness of even
a snapshot of precise location data should not be
understated.” United States v. Smith, 110 F.4th 817, 833
(5th Cir. 2024). The geofence in this case centered on “a
busy part of the Richmond metro area” between 3:50 and
5:50 p.m., when many people are leaving work or
school—and of course, it had no geographic boundaries
at Step 2. Chatrie, 590 F. Supp. 3d at 925; see id. at 919.
Two hours of Location History for accounts passing
through that geofence could enable police to tour a
person’s home, capture their romantic rendezvous, or
accompany them to church.
This case presents textbook examples of how police
access to this digital information can invade the
privacies of innocent users. At the suppression hearing,
Chatrie’s counsel demonstrated that the anonymized
Step 2 data produced in response to this geofence
warrant tracked three innocent users to or from private
spaces, including residences, a school, and a hospital. Id.
at 923–24. Chatrie’s expert showed how this information,
when combined with publicly available information,
allowed him to easily deduce those individuals’
identities. Id.8
8
Whether the Location History collected here placed Chatrie
himself inside a constitutionally protected space is beside the point.
“In Carpenter, the Supreme Court’s analysis of whether the
55a
Some of my colleagues believe that because a two-hour
snippet of Location History is too short to “reveal
intimate details through habits and patterns,” like the
aerial surveillance footage in Leaders of a Beautiful
Struggle v. Baltimore Police Department, 2 F.4th 330,
341 (4th Cir. 2021), it cannot reveal intimate details at
all. See Judge Richardson Concurring Op., infra, at 140
n.19. But pattern-based deductions are not the only way
Government’s access of the defendant’s CSLI impeded his
reasonable expectation of privacy was not based on a review of the
specific results of the search in that case.” United States v. Smith,
110 F.4th 817, 834 n.8 (5th Cir. 2024) (citing Carpenter, 585 U.S. at
309–13, 138 S.Ct. 2206). Instead, “[t]he question was whether the
technology utilized by law enforcement had the capability of
providing data that offered ‘an all-encompassing record of [a
person’s] whereabouts,’ regardless of whether that person actually
entered spaces that are traditionally considered protected under
the Fourth Amendment.” Id. (quoting Carpenter, 585 U.S. at 311,
138 S.Ct. 2206).
Similarly, Kyllo rejected the argument that the search of heat
waves emanating from a home did not implicate the Fourth
Amendment if the search did not reveal intimate details. That
argument, Justice Scalia explained, was not only “wrong in
principle,” but also “impractical” because “no police officer would be
able to know in advance whether his through-the-wall surveillance
picks up ‘intimate’ details—and thus would be unable to know in
advance whether it is constitutional.” Kyllo, 533 U.S. at 38–39, 121
S.Ct. 2038. Likewise, when police drew up a geofence that included
private spaces, they could not predict whether Chatrie would be
shown to have entered those spaces. The Government cannot
circumvent the Constitution merely because, by sheer luck, its
target did not stray from the safe zone. See Arizona v. Hicks, 480
U.S. 321, 325, 107 S.Ct. 1149, 94 L.Ed.2d 347 (1987) (“A search is a
search, even if it happens to disclose nothing but the bottom of a
turntable.”).
56a
to uncover intimate personal details.9 Another way is to
use a surveillance technology that can follow subjects
through walls. See Kyllo, 533 U.S. at 37–39, 121 S.Ct.
2038. The aerial surveillance program at issue in
Beautiful Struggle tracked only public movements, so
our short-term–long-term distinction made sense; it
takes a lot of grainy aerial footage to deduce intimate
personal details.10 Location History’s accuracy—not to
mention its vast retrospective scope—makes it a much
more potent tool.
A few of my colleagues claim that “[a] record of a
person’s single, brief trip is no more revealing than his
bank records or telephone call logs.” Judge Richardson
Concurring Op., infra, at 140. Respectfully, that is
wrong on multiple accounts. Most obviously, it flat-out
ignores the public surveillance doctrine. Tracking a
person’s “single, brief trip” on public thoroughfares (as
in Knotts) is not a search; but tracking even an object’s
trip in and out of a private space (as in Karo) is a search.
Compare Knotts, 460 U.S. at 281, 103 S.Ct. 1081 with
Karo, 468 U.S. at 714–16, 104 S.Ct. 3296. Location
History is capable of tracking people in and out of private
spaces, with even greater precision than CSLI or the
9
Indeed, Carpenter made no mention of habits or patterns in
discussing the capabilities of CSLI.
10
The weeks-long aerial surveillance program at issue in Beautiful
Struggle monitored only public spaces during the day, gathered
hours-long chunks of image data in which people appeared as blurry
collections of pixels, and stored that data for forty-five days.
Beautiful Struggle, 2 F.4th at 334, 341–42. As a result, the
Government had to decipher individuals’ identities from several
pieces of captured data. Id. at 344–45.
57a
beeper in Karo. More tellingly, Carpenter expressly
recognized that the deeply revealing nature of “cell
phone location records” puts them in a “qualitatively
different category” from “telephone numbers and bank
records.” Carpenter, 585 U.S. at 309, 138 S.Ct. 2206.
Carpenter’s observation about CSLI is doubly true of
Location History.
In light of the intimately revealing nature of Location
History data, the span of time it covers is of little
importance to the Fourth Amendment search analysis.
The Government in Carpenter requested CSLI
spanning both seven- and 152-day periods, which
revealed, respectively, two and 127 days of data. Id. at
302, 138 S.Ct. 2206. But Carpenter ultimately held that
accessing the shorter span of data was enough to
constitute a Fourth Amendment search. Id. at 310 n.3,
138 S.Ct. 2206. The Court’s intimacy analysis drew on
Justice Sotomayor’s concurrence in Jones, which argued
that even short-term GPS tracking violates reasonable
expectations of privacy. See id. at 311, 138 S.Ct. 2206
(citing Jones, 565 U.S. at 415, 132 S.Ct. 945 (Sotomayor,
J., concurring)).
Moreover, Carpenter focused on the nature of the search
technology employed, not the duration of the particular
search at bar. Even though the Government only
accessed discrete segments of Carpenter’s CSLI, the
Court stressed repeatedly that carriers collect and store
CSLI for “years.” Id. at 312, 313, 315, 319, 138 S.Ct. 2206.
Location History collects even more (and more precise)
location data, and stores it indefinitely. Applying
Carpenter’s logic, police use of a technology whose very
purpose is to generate a dossier of intimately revealing
58a
location data traceable to individuals is a search—even
if only a snippet is ultimately obtained.
At bottom, focusing on the duration of the geofence
employed in this particular case “overlooks the critical
issue”: that Location History “is an entirely different
species of business record[,] something that implicates
basic Fourth Amendment concerns about arbitrary
government power much more directly than corporate
tax or payroll ledgers.” Id. at 318, 138 S.Ct. 2206. There
can be no doubt that even a small amount of such data
“provides an intimate window into a person’s life.” Id. at
311, 138 S.Ct. 2206.
D.
Carpenter also found it significant that CSLI searches
are “easy, cheap, and efficient compared to traditional
investigative tools.” Carpenter, 585 U.S. at 311, 138 S.Ct.
2206. That concern echoes the Jones concurrences,
which warned that low-cost surveillance technologies
could lead to more surveillance and less accountability.
Justice Sotomayor’s concurrence noted that “because
GPS monitoring is cheap in comparison to conventional
surveillance techniques and, by design, proceeds
surreptitiously, it evades the ordinary checks that
constrain abusive law enforcement practices: ‘limited
police resources and community hostility.’” Jones, 565
U.S. at 415–16, 132 S.Ct. 945 (Sotomayor, J., concurring)
(quoting Lidster, 540 U.S. at 426, 124 S.Ct. 885). And
Justice Alito added that GPS technology “makes longterm monitoring”—which was traditionally “difficult
and costly and therefore rarely undertaken”—
59a
“relatively easy and cheap.” Id. at 429, 132 S.Ct. 945
(Alito, J., concurring in the judgment).
Location History is like the GPS monitoring in Jones,
only cheaper and more intrusive. Scholars have
estimated that “tracking location by cell phone,” as
police did in Carpenter, “is almost twice as cheap as GPS
tracking,” which in turn is “twenty-eight times cheaper
than covert pursuit.” Ohm, supra n.4, at 369 (citing
Kevin S. Bankston & Ashkan Soltani, Tiny Constables
and the Cost of Surveillance: Making Cents Out of
United States v. Jones, 123 Yale L.J. Online 335, 354
(2014)). Location History tracking is likely even cheaper.
“With just the click of a button,” Google—at the
Government’s request—“can access [its] deep
repository of historical location information at
practically no expense” to the Government. Carpenter,
585 U.S. at 311, 138 S.Ct. 2206. And unlike the tracking
device in
This text is long and has been trimmed here. Open the source document for the complete record.
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.