Every Documented Wrongful Facial Recognition Arrest Involved a Black Person. Here's the Real Legal Landscape.

No federal law exists, so states and cities are writing the actual rules

There is no federal law governing police use of facial recognition technology in the United States. As of late 2025, 15 states, Washington, Oregon, Montana, Utah, Colorado, Minnesota, Illinois, Alabama, Virginia, Maryland, New Jersey, Massachusetts, New Hampshire, Vermont, and Maine, had some form of legislation restricting police use, and at least 16 cities have gone further, banning the technology for police use entirely. San Francisco led the city-ban trend in 2019, followed by Boston, Oakland, and others; Milwaukee became the latest to join, voluntarily banning all police facial recognition in February 2026 following public outcry.

Seven states, Colorado, Maryland, Maine, Virginia, Washington, Montana, and Alabama, have adopted what's sometimes called a "sole basis" prohibition, meaning police cannot make an arrest based on a facial recognition match alone, without independent corroborating evidence. Alabama's law, while enacted, doesn't actually take effect until 2027, illustrating how much lag time exists even after a state formally legislates on this issue.

Every publicly documented wrongful arrest has involved the same racial pattern

According to compiled tracking from State of Surveillance, every publicly known wrongful arrest resulting from facial recognition technology has involved a Black person. In 2019 and 2020, three Black men were separately accused of and jailed for crimes they didn't commit after police used facial recognition to falsely identify them. A new, confirmed case occurred in August 2025, when NYPD falsely arrested Trevis Williams based on a wrongful facial recognition match. The ACLU has separately documented that facial recognition technology performs measurably worse at identifying Black individuals, a well-established technical limitation, not just an anecdotal pattern.

City-level bans have a real, documented loophole

• A Washington Post investigation published in May 2024 found that police departments in cities with facial recognition bans, including Austin and San Francisco, were circumventing their own local prohibitions by asking other law enforcement agencies without such restrictions to run searches on their behalf
• This specific loophole, a city's police department calling a nearby county sheriff's office that has no ban to run the search instead, is a key reason advocates argue state-level legislation matters more than city bans alone, since state law can close inter-jurisdictional workarounds that city ordinances by themselves cannot
• Advocates specifically note that no wrongful facial recognition arrest has been publicly reported in a city with an active, functioning ban, framing this as the clearest practical argument for bans working as intended when not circumvented
• State of Surveillance's tracking specifically flags New York and Virginia as two states where 2025-2026 legislative activity is most active, with New York's Facial Recognition Technology Study Act (S3699) passing the State Senate but not yet becoming law as of the most recent tracking

Most existing state laws only cover a fraction of how the technology is actually used

A significant, often overlooked limitation of existing state laws: most only restrict law enforcement use specifically. They generally don't cover facial recognition used commercially, meaning a person can still be tracked by facial recognition technology at a concert, in a retail store, or in other private commercial settings even in states with relatively strong police-focused restrictions. Illinois's Biometric Information Privacy Act (BIPA) is a notable exception, since it covers commercial use broadly and has produced major settlements, including a $650 million Facebook settlement and a $1.4 billion Google settlement, over alleged unauthorized biometric data collection.

The core disagreement

Law enforcement and technology advocates generally argue facial recognition provides a genuinely valuable investigative tool for identifying suspects and solving crimes faster than traditional methods, and that appropriate safeguards, like sole-basis prohibitions and accuracy standards, adequately address the technology's real limitations without banning a useful tool outright. Privacy and civil rights advocates generally argue the documented, consistent racial bias in wrongful arrests, combined with real evidence that city-level bans are being circumvented through inter-agency requests, shows current safeguards are insufficient, and argue for stronger, harder-to-circumvent restrictions or outright bans rather than narrower accuracy and disclosure rules alone. Both sides broadly agree that the current legal landscape, no federal law, a genuinely uneven patchwork of state and city rules, and a real technical loophole in how bans can be circumvented, is inadequate as a long-term, stable framework, even though they disagree sharply on what a better framework should actually require.