Two courts have looked at automated license plate readers (ALPRs) and reached opposite conclusions. In United States v. Kyle, a federal court in Oklahoma held that a warrantless ALPR search violated the Fourth Amendment and described the technology as "a type of indiscriminate mass surveillance," according to the Electronic Frontier Foundation. A California trial court dismissed SIREN v. San Jose, which challenged San Jose's warrantless searches of its own ALPR database. EFF, which represents the plaintiffs with the ACLU of Northern California, says it will appeal.
The split matters beyond these two cases. It shows that how much protection a driver has depends on which courthouse hears the case. That is a poor basis for rules governing a technology deployed nationwide.
The strongest case for the police
The argument for ALPRs deserves a fair statement. A license plate is displayed to the public by legal requirement. Officers may lawfully read plates by eye, and a camera that reads them faster does the same thing. ALPR data helps locate stolen vehicles, abducted children and fleeing suspects, and a blanket warrant requirement for every query could slow urgent investigations. The California court appears to have accepted a version of this reasoning. EFF says it relied on an appellate decision about individual streetlight cameras and concluded that people have no reasonable expectation of privacy while driving on public roads.
What the Oklahoma facts show
The Oklahoma record shows why that analogy fails. In the opinion and order, filed October 1, 2026, the court describes a Tulsa County deputy who followed a Mazda with California plates "for no apparent reason other than the fact that it had a California license plate." He then queried the Flock system. The court notes that ALPR networks gather "historical data about a vehicle's whereabouts on public roadways for use by law enforcement across the country," and that the deputy could see at least 30 days of location history in many states. The query showed the car's trip from Oklahoma toward the California-Arizona border and back toward Missouri.
That is not a camera reading a plate at one intersection. It is a retrospective map of a person's movements, retrieved with no suspicion beyond an out-of-state tag. The court granted the motion to suppress.
Why aggregation changes the analysis
The Supreme Court has already addressed the underlying logic. In Carpenter v. United States (2018), it held that obtaining historical cell-site location records is a Fourth Amendment search, even though each data point was generated in public view. EFF also points to Chatrie v. United States (2026), which it describes as recognizing an expectation of privacy in location data that reveals physical movements. Under that line of cases, a single observation is not the issue. The issue is the compiled record.
Scale matters too. EFF reports that San Jose's system has nearly 500 cameras photographing every passing car, with data kept for a month or longer. That is a standing database of ordinary residents' movements, searchable without a warrant.
Governance failures reinforce the point
The problem is not only constitutional. In a separate October 1 post, EFF reported that, according to audit logs obtained by the Point Reyes Light, the Marin County Sheriff's Office shared Flock ALPR data with out-of-state and federal agencies in late 2024, including 254,131 searches in November 2024 alone. EFF says this violates California's SB 34. A 2022 settlement in Lagleva v. Marin County Sheriff had already required the Sheriff to stop sharing with agencies outside California. The sheriff's office says it disabled nationwide access, but EFF says it has not explained how outside agencies still got in.
This undercuts the standard reassurance that agency policies and vendor settings are enough. When one sheriff's office with a prior settlement still cannot control who queries its data, a system built on voluntary compliance is not working.
A proportionate path
A pro-innovation position does not require rejecting the technology. It requires rules that keep the technology's benefits while cutting off its worst uses. Legislatures can do that more reliably than appellate litigation, which takes years and produces outcomes that vary by jurisdiction. Sensible elements include:
- Short default retention. Data on vehicles not tied to an investigation should be deleted within days, not months.
- A documented, case-linked justification for every query, with a warrant required for lookback searches beyond a short window.
- A ban on sharing with agencies outside the jurisdiction unless a warrant or statute authorizes it.
- Audit logs that are available to oversight bodies and, in aggregate, to the public.
- Vendor accountability when sharing settings fail.
These rules would keep the emergency uses, such as stolen cars and Amber Alerts, while ending the practice of searching a network of cameras to profile drivers on a hunch.
What to watch
The Oklahoma ruling is a single district court decision, and the California dismissal is a trial-level ruling that EFF says it will appeal. Neither settles the law. But the two together tell lawmakers that the constitutional question is open and that waiting for courts to answer it leaves both civil liberties and police practice uncertain. Clear statutory limits would give law enforcement predictable rules and give residents real protection.