Shield Public Safety Training · Police Technology Case Law Center

People v. Seymour

536 P.3d 1260 (Colo. 2023) · 2023 CO 53

A landmark Colorado Supreme Court decision addressing reverse-keyword warrants, privacy in Google search history, expressive interests, probable cause, particularity, and good-faith reliance when police ask a search provider to identify unknown users based on what they searched.

CourtColorado Supreme Court
DecisionOctober 16, 2023
DocketNo. 23SA12
OpinionJustice William W. Hood III
InvestigationFatal arson and homicide
TechnologyGoogle reverse-keyword warrant
DataSearch queries and IP addresses
Primary IssuesPrivacy, probable cause and particularity
Additional ConcernFreedom of expression
DispositionSuppression denied; case remanded

Executive Summary

The Case in One Paragraph

Denver police investigated an August 2020 arson that killed five people. After extensive conventional investigation failed to identify a suspect, detectives theorized that the perpetrators may have searched online for the targeted residential address before the crime. Police obtained a reverse-keyword warrant directing Google to identify users associated with searches for specified address-related terms during roughly the two weeks preceding the fire. Google ultimately disclosed records that included five Colorado IP addresses associated with responsive searches, and further investigation helped identify Gavin Seymour and others. Seymour moved to suppress the resulting evidence. The Colorado Supreme Court held that Seymour had a constitutionally protected privacy interest in his Google search history under the Colorado Constitution and a possessory interest in that information under both the Colorado and federal constitutions. The court also recognized that search history implicates freedom of expression and therefore requires particularly careful application of constitutional protections. It held that the warrant adequately particularized the place to be searched and things to be seized. The court assumed without deciding that individualized probable cause was required and absent, but nevertheless refused suppression because investigators obtained and executed the novel warrant in objectively reasonable good faith.

Core Rule Reverse-keyword warrants implicate significant constitutional interests because internet search history can reveal private thought and expressive activity. Seymour did not categorically approve or prohibit such warrants: it upheld denial of suppression because the warrant was sufficiently particular and investigators acted in good faith, while assuming without deciding that individualized probable cause was constitutionally required and lacking.

Key Holdings at a Glance

Search History Is Constitutionally Sensitive The court recognized a protected privacy interest under the Colorado Constitution in Google search history even when initially linked only to an IP address rather than a person's name.
Possessory Interests Also Applied Government acquisition of Seymour's Google search-history information meaningfully interfered with his possessory interest in that data under both federal and Colorado constitutional principles.
Expression Heightened the Concern Internet searches can reveal interests, questions, beliefs, concerns, and intellectual activity, implicating freedom-of-expression protections.
Particularity Was Satisfied The warrant specified the Google data to be searched and nine search terms or keywords sufficiently to constrain the search.
Probable Cause Was Unresolved The majority assumed without deciding that individualized probable cause was required and absent rather than establishing a categorical rule for all keyword warrants.
Good Faith Prevented Suppression The technique was novel, investigators worked with Google and prosecutors, modified the warrant, obtained judicial approval, and acted reasonably under unsettled law.

The Fatal-Arson Investigation

On August 5, 2020, a fire at a Denver residence killed five people. Investigators concluded that an accelerant had been used, and nearby home-surveillance footage showed three masked people carrying what appeared to be a gasoline container near the residence when the fire began.

Denver police conducted an extensive investigation. Officers interviewed neighbors, reviewed surveillance footage, examined activity at nearby gas stations, obtained numerous search warrants, and pursued other leads. More than two months passed without identification of a suspect.

Investigators eventually theorized that whoever intentionally targeted the residence might have researched the property beforehand or used Google to obtain directions to the address.

Investigative Context Seymour involved a reverse-keyword warrant used only after investigators had exhausted substantial conventional investigative leads. That history became relevant to the court's assessment of reasonableness and good faith.

What Is a Reverse-Keyword Warrant?

A traditional digital warrant usually starts with an identified suspect, account, device, email address, or telephone number. Investigators establish probable cause concerning that target and seek specified records.

A reverse-keyword warrant works in the opposite direction. Investigators begin with words, phrases, an address, product name, victim name, or another search term believed to be connected with criminal activity. They then ask a search provider to identify users who previously searched for that term.

In Seymour, investigators sought information associated with searches relating to the address where the fatal fire occurred during a defined period before the crime.

Google initially resisted aspects of the requested process, leading investigators to revise the warrant language. The final warrant identified nine specific keyword variations connected with the address.

Reverse Identification A keyword warrant uses expressive conduct—the content of an internet search—to identify a person who is not yet known to investigators. That makes it analytically different from obtaining the search history of an already identified suspect.

Privacy in Google Search History

The Colorado Supreme Court recognized that internet search history can reveal extraordinarily personal information. Search queries may disclose health concerns, political interests, religious questions, intimate relationships, financial difficulties, personal fears, intellectual curiosity, and countless other subjects a person may never communicate publicly.

The court concluded that Seymour had a constitutionally protected privacy interest in his Google search history under Article II, section 7 of the Colorado Constitution, even though the initial data was associated with an IP address rather than Seymour's name.

The court also concluded that the government meaningfully interfered with Seymour's possessory interest in his search-history data, implicating search-and-seizure protections under both the Fourth Amendment and Colorado Constitution.

Digital Privacy Principle Anonymization does not necessarily eliminate constitutional protection. Search history linked initially only to an IP address can still implicate protected privacy and possessory interests.

Search History and Freedom of Expression

Seymour is particularly significant because the court did not treat internet search history merely as transactional data.

Online searches are part of how people seek information, investigate ideas, read about controversial subjects, explore political or religious questions, research medical conditions, and engage privately with information.

The court therefore concluded that Seymour's search history implicated freedom-of-expression interests. When search-and-seizure authority reaches expressive material, constitutional protections must be applied with what Supreme Court precedent has described as heightened exactitude.

Expressive Data A keyword warrant can identify someone not because of where the person went or whom the person contacted, but because of what the person thought important enough to search. That expressive dimension materially increases the constitutional sensitivity of the technique.

Why the Warrant Satisfied Particularity

Seymour argued that the warrant failed the Fourth Amendment's particularity requirement because the true place searched was effectively Google's enormous database.

The court rejected that categorical approach. It concluded that although Google's infrastructure may contain an immense universe of information, the warrant's specific search parameters substantially narrowed what Google was authorized to identify.

The final warrant was limited to nine specified keyword variations and a defined period. Those parameters sufficiently described the information sought.

Google ultimately produced more records than exactly matched the specified keywords because some searches contained the terms along with additional words. The court held that Google's overproduction did not retroactively render the warrant insufficiently particular. The constitutional inquiry focused on what the warrant authorized, not merely what the provider chose to disclose.

Particularity Holding The warrant adequately described the place to be searched and the things to be seized because it confined Google's search through specified keyword terms and temporal parameters rather than authorizing an unrestricted search for any information investigators considered relevant.

The Harder Question: Individualized Probable Cause

Particularity was only part of the constitutional inquiry. Seymour also argued that police lacked probable cause to search his Google records because investigators had no evidence connecting him—or any particular Google user—to the crime before the keyword search occurred.

The majority did not definitively resolve whether every reverse-keyword warrant requires individualized probable cause as to the persons whose search information is examined.

Instead, the court assumed without deciding that individualized probable cause was required and that its absence rendered the warrant constitutionally defective. It then moved to the separate question of suppression.

Do Not Overstate the Holding Seymour did not definitively hold that the warrant lacked probable cause under a categorical rule governing reverse-keyword warrants. The majority assumed that proposition for purposes of deciding whether suppression was required.

The court nevertheless emphasized an important nexus limitation: the government cannot justify a keyword warrant merely because a word or location is searchable. Investigators need a meaningful connection between the search term and the suspected criminal activity.

Nexus Matters The court distinguished searches tied to a specific crime location from hypothetical attempts to identify everyone who searched a church, doctor's office, or other sensitive location without evidence connecting that place or term to criminal activity.

Why the Good-Faith Exception Applied

Even assuming the warrant was defective for lack of individualized probable cause, the Colorado Supreme Court concluded that suppression would not serve the purpose of the exclusionary rule.

Reverse-keyword warrants presented a novel constitutional issue. At the time investigators acted, no state supreme court or federal appellate court had addressed their constitutionality.

The record showed that investigators did not simply demand data informally. They worked through legal process, communicated with Google concerning the warrant's wording, revised the request when Google objected, obtained judicial authorization, and executed the search pursuant to the approved warrant.

The court concluded that law enforcement had acted objectively reasonably in navigating a new investigative technique for which controlling precedent did not yet exist.

Remedial Holding Because investigators reasonably relied on judicial authorization while using a novel technique in an unsettled legal environment, suppressing the resulting evidence would not meaningfully deter police misconduct.
Future Good Faith The court expressly refused to give blanket approval to reverse-keyword warrants. As judicial guidance develops, agencies cannot assume that the same novelty-based good-faith rationale will protect future searches.

The Dissent's General-Warrant Concern

Justice Márquez, joined by Justice Samour, strongly disagreed with the majority's treatment of the warrant.

The dissent characterized reverse-keyword searching as a digital dragnet because investigators ask a provider to search across a vast universe of users' private search histories in order to identify people who typed particular terms.

From that perspective, the lack of an identified suspect and absence of individualized probable cause made the process resemble the general warrants that the Fourth Amendment was adopted to prohibit.

The dissent also emphasized the expressive nature of internet searches, reasoning that keyword warrants can expose what amounts to a record of users' private intellectual activity.

Unresolved Tension Seymour exposes the central constitutional debate surrounding keyword warrants: whether sufficiently specific search terms can particularize a reverse search, or whether searching an immense population's private queries to discover a suspect is structurally incompatible with individualized probable cause.

What Seymour Does—and Does Not—Establish

  • Seymour does not categorically approve reverse-keyword warrants.
  • Seymour does not categorically prohibit reverse-keyword warrants.
  • The Colorado Supreme Court recognized substantial privacy interests in Google search history.
  • The court recognized a possessory interest in search-history data under both federal and Colorado constitutional principles.
  • The court held that internet search history implicates expressive interests.
  • The warrant satisfied the particularity requirement because the search terms and scope were sufficiently defined.
  • The majority assumed without deciding that individualized probable cause was required and absent.
  • The evidence was admitted because investigators acted in objectively reasonable good faith.
  • The decision emphasizes that a search term must have a meaningful nexus to criminal activity.
  • Provider overproduction beyond the warrant's terms does not automatically invalidate an otherwise particular warrant.
  • The court expressly left room for future courts to restrict reverse-keyword searching if technological expansion creates broader privacy threats.

Seymour Compared with Related Digital-Search Decisions

Case Technology or Evidence Central Rule
People v. Seymour Reverse-keyword warrant and Google search history Search history receives constitutional protection; warrant was particular; individualized probable cause assumed lacking; good faith prevented suppression.
Riley v. California Cellphone contents Digital devices contain vast quantities of deeply private information and generally require a warrant to search incident to arrest.
Carpenter v. United States Historical cellphone location information Digital aggregation can create privacy interests despite records being held by a third-party provider.
United States v. Smith Geofence reverse-location warrant Fifth Circuit found reverse-location warrants unconstitutional general warrants but applied good faith.
People v. Meza California geofence warrant Warrant was overbroad and insufficiently particular; evidence survived under good faith.
Price v. Superior Court Narrow California geofence warrant Tightly constrained reverse-location warrant satisfied probable cause and particularity.
United States v. Leon Reliance on defective warrant Objectively reasonable reliance on judicial authorization may prevent suppression.

Agency Operations Checklist

  1. Use reverse-keyword searching only for a clearly defined investigative need. Document why conventional techniques have failed or are unlikely to identify the unknown suspect.
  2. Establish a strong nexus between the keyword and the crime. Explain why the perpetrator likely searched the exact term or closely related terms.
  3. Avoid generic or high-volume terms. Broad terms increase innocent-user exposure and weaken the connection between a responsive search and criminal activity.
  4. Define every keyword expressly. Do not allow the provider or investigator to improvise additional terms after judicial authorization.
  5. Use the shortest defensible time period. Tie the search window to the known preparation period or offense timeline.
  6. Address expressive interests. Consider whether the requested terms concern religion, politics, health, sexuality, journalism, legal research, advocacy, or other sensitive expressive activity.
  7. Explain provider mechanics. Describe whether Google searches accounts, queries, IP addresses, device identifiers, or another dataset and what investigators will actually receive.
  8. Minimize identification of uninvolved users. Consider staged disclosure, anonymization, judicial reauthorization, or other procedures that reduce unnecessary exposure.
  9. Consider renewed judicial review before unmasking users. Especially where many responsive users are returned, additional probable-cause review can strengthen the constitutional process.
  10. Consult current appellate authority. Keyword-warrant law remains developing and may differ significantly by jurisdiction.
  11. Preserve the complete warrant process. Retain draft warrants, provider objections, revisions, prosecutor communications, judicial orders, returns, and narrowing decisions.
  12. Corroborate independently. A responsive keyword search is an investigative lead, not proof that the searcher committed the crime.

Litigation and Review Checklist

  • Identify the exact search terms authorized by the warrant.
  • Determine the factual nexus between each term and the offense.
  • Identify the time period searched and explain its relationship to the crime.
  • Determine what Google or another provider actually searched internally.
  • Determine how many users or queries were potentially examined.
  • Identify how many responsive accounts, IP addresses, or identifiers were returned.
  • Separate provider overproduction from the scope actually authorized by the warrant.
  • Analyze privacy and possessory interests independently.
  • Determine whether expressive or First Amendment interests trigger heightened constitutional concerns.
  • Analyze particularity separately from probable cause.
  • Determine whether probable cause was individualized or merely based on a generalized inference that the perpetrator might have searched the term.
  • Identify whether additional judicial approval was obtained before users were identified.
  • Analyze good faith separately from constitutional validity.
  • Determine what law existed when investigators sought and executed the warrant.
  • Trace downstream warrants and evidence to determine how heavily they depended on the keyword search.

Frequently Asked Questions

What is a reverse-keyword warrant?

It is a warrant directing a search provider to identify users who searched specified words, phrases, addresses, or other terms during a defined period even though investigators do not yet know who the suspects are.

Did Seymour hold that people have privacy rights in Google search history?

Yes under the Colorado Constitution. The court also held that Seymour had a constitutionally protected possessory interest in his search-history data under both Colorado and federal constitutional principles.

Did the court find the warrant insufficiently particular?

No. The court held that the warrant adequately described the data to be searched and the specified search terms.

Did the court hold that the warrant lacked probable cause?

Not categorically. The majority assumed without deciding that individualized probable cause was constitutionally required and absent, then resolved the suppression issue through good faith.

Why did search history implicate freedom of expression?

Internet searches reveal what information people seek and can disclose private intellectual, religious, political, medical, personal, and other expressive interests.

What information did Google provide?

The investigation ultimately received information including Colorado IP addresses associated with responsive searches. Further legal process and investigation were then used to identify suspects.

Did the court approve all reverse-keyword warrants?

No. The majority expressly declined to make a broad proclamation regarding their validity and emphasized that future constitutional analysis would remain fact dependent.

Why was the evidence not suppressed?

Investigators were using a novel technique in an unsettled area of law, revised the warrant in response to provider concerns, obtained judicial approval, and acted in objectively reasonable good faith.

Can agencies rely on Seymour's good-faith ruling for future keyword warrants?

Not automatically. Good faith depends on the law existing when investigators act. As appellate guidance develops, techniques previously treated as novel become subject to clearer constitutional expectations.

Primary Authorities and Related Law

People v. Seymour, 536 P.3d 1260 (Colo. 2023), 2023 CO 53
Colorado Supreme Court decision addressing reverse-keyword warrants, Google search-history privacy, expression, particularity, probable cause, and good faith.
Read People v. Seymour
Riley v. California, 573 U.S. 373 (2014)
Supreme Court decision recognizing the exceptional quantity and sensitivity of personal information stored in modern digital devices.
Read Riley v. California
Carpenter v. United States, 585 U.S. 296 (2018)
Supreme Court decision recognizing significant Fourth Amendment protection for stored digital location information held by third parties.
Read Carpenter v. United States
Stanford v. Texas, 379 U.S. 476 (1965)
Supreme Court authority emphasizing particularly exacting application of warrant requirements when government searches implicate expressive material.
Read Stanford v. Texas
United States v. Leon, 468 U.S. 897 (1984)
Supreme Court decision establishing the good-faith exception for objectively reasonable reliance on judicially issued warrants.
Read United States v. Leon

Final Assessment

People v. Seymour is an important marker in the evolution of digital-search law because reverse-keyword warrants move beyond surveillance of where a person went and into surveillance of what a person sought to know. Search queries can reveal thoughts, interests, fears, research, beliefs, medical concerns, political questions, religious inquiry, and countless other forms of private intellectual activity.

The Colorado Supreme Court therefore treated the data as constitutionally significant. At the same time, the majority declined to establish a categorical rule either permitting or prohibiting keyword warrants. It held that the warrant was sufficiently particular, assumed a probable-cause defect for purposes of analysis, and preserved the evidence because investigators reasonably relied on judicial process while navigating a technique no appellate court had yet squarely addressed.

For agencies, the decision is not an invitation to treat keyword searching as routine. It is a warning that reverse-search techniques involving expressive data require unusually careful nexus, probable-cause, particularity, minimization, and judicial-review analysis.

Shield Practice Rule Treat reverse-keyword warrants as exceptionally sensitive digital searches. Establish a concrete nexus between the exact search terms and the offense, use the narrowest supportable keywords and time period, minimize exposure of uninvolved users, consider renewed judicial approval before identification, address expressive interests explicitly, and do not rely on the novelty-based good faith that preserved the evidence in Seymour.

Shield Public Safety Training · Police Technology Case Law Center

This monograph is provided for training and general informational purposes. It is not legal advice and does not replace review of complete opinions, current statutes, court rules, controlling federal and state authority, provider requirements, agency policy, technical documentation, or consultation with prosecutors and agency counsel.

© 2026 Shield Public Safety Training. All rights reserved. Reviewed August 30, 2026.

```