When a Warrant Looks Like a General Warrant

Once access to historical Location History is recognized as a search, the central legal problem shifts to the warrant’s design. This chapter explains how probable cause and particularity apply when the government starts with a place and time rather than a known person, and when later requests become more intrusive than the first. The practical judgment is stage-specific: ask what place, person, or thing was searched, what fair probability supported that request when authorization was sought, and who controlled the decision to escalate.

Probable cause is not a database-wide hunch

The Fourth Amendment generally requires a warrant supported by probable cause when officers search for evidence of a crime. Probable cause is not mathematical certainty. It asks whether the facts establish a fair probability that evidence will be found in the place searched and that the evidence will aid the investigation. Particularity performs a different function. It limits the scope of the search by requiring the warrant to identify the place to be searched and the persons or things to be seized with enough precision to prevent a wide-ranging exploratory search.

Those controls fit comfortably when police know which house, account, or device they are investigating. A geofence reverses the starting point. The officers know where the crime occurred but do not know which person’s records will reveal evidence. The question becomes whether probable cause to believe that evidence exists somewhere in a provider’s database also supplies probable cause to search every user who might fit a place-time description.

The Fifth Circuit answered no in United States v. Smith. The case involved the robbery of a Postal Service route driver in Mississippi. Investigators had a crime scene, video, vehicles, and an indication that the assailant might have been using a cell phone, but no identified suspect. They obtained a geofence warrant from Google. The first stage produced anonymous device identifiers associated with the specified area and time. Police then selected devices for broader movement data and eventually requested account information.

The court treated the first stage as constitutionally significant because Google could not produce the requested result without searching its entire Location History repository. In its description, the government was not asking for a known user’s record. It was compelling a provider to search across a vast population and return whoever appeared in the geographic and temporal box. The Fifth Circuit reasoned that Google acted as an instrument or agent of the government because it performed the search in response to legal compulsion.

That reasoning led to its general-warrant conclusion. A traditional general warrant specified an offense but left the executing officials to decide which people to arrest or which places to search. The Fifth Circuit saw the geofence as a modern version of that problem: the warrant named a crime, a location, and a time, but not a particular suspect or a defined group for which individualized facts existed. It allowed the search itself to discover who might be relevant.

The difference between a search and its result is central to this argument. Suppose a broad search eventually produces only two names, and both later appear strongly connected to the crime. That outcome may make the investigation look narrow in retrospect. But the initial search was performed before the officers knew which two people would emerge.