First man wrongfully arrested because of facial recognition testifies as California weighs new bills
In 2020, Robert Williams was arrested for allegedly stealing thousands of dollars of watches. Detroit police had matched grainy surveillance footage of the crime to Williams’ driver’s license photo using a facial recognition service. But Williams wasn’t the robber. At the time of the robbery, he was driving home from work.
Williams’ arrest was the first documented case of someone being wrongfully detained based on facial recognition technology, which is used by police departments and government agencies across the US.
On Tuesday, Williams testified before the California legislature, which is considering several proposals about the use of the technology by police.
“It’s dangerous when it works and even more dangerous when it doesn’t work,” Williams said.
“In my case, Detroit police were supposed to treat face recognition matches as an investigative lead, not as the only proof they need to charge someone with a crime,” Williams wrote in a letter to the California assembly’s public safety committee. “They should have collected corroborating evidence such as an eyewitness identification, cell phone location data or a fingerprint. They had none of that – just an out-offocus image of a large Black man in a baseball cap that a faulty algorithm had determined was me.”
Williams’ testimony comes as the California legislature is mulling two proposals. The first bill would place a moratorium on the use of facial recognition in police body cameras, while the second identifies circumstances in which police would both be permitted to use and restricted from using facial recognition systems.
The bills reflect an ongoing debate and tension over the use of surveillance systems in California. In San Francisco, for example, the board of supervisors in 2022 voted to allow the city’s police department to access live footage from private surveillance cameras. Just four years earlier, that very same board had voted to ban police and other city agencies from using facial recognition.
Several studies show that facial recognition systems regularly misidentify Black and brown people, posing a particular threat to communities that have already been disproportionately targeted by police and surveillance systems.
It’s in the spirit of limiting the technology that AB 1034, which was sponsored by assembly members Lori Wilson and Akilah Weber, was introduced. If signed into law, the bill would extend a three-year moratorium on the use of facial recognition in police body cameras for another 10 years.
The bill received broad support from a coalition of civil liberty organizations, led by the American Civil Liberties Union. ACLU of Northern California technology and civil liberties attorney Matt Cagle said the organization sees the proposal as just a first step. “The ACLU and our partners do not think facial recognition ever would be appropriate on these cameras, to be clear.”
“This bill makes sure that body cameras, which were promised as a tool for police accountability for transparency [around] police conduct, can’t be transformed and exploited for face surveillance,” said Cagle.
Simultaneously, lawmakers are considering another bill, AB 642, introduced by Philip Ting, that would provide guidelines to police and government agencies about when they could deploy facial recognition software.
Ting said his bill is simply meant to impose basic guardrails, and doesn’t prevent local authorities from stronger action.
But some of the committee members, including Wilson, who represents Suisun City, voiced their reservations over whether the bill expands the use of facial recognition in the state. Of particular concern is an amendment that could allow police to try to match images of people’s faces against the California driver’s license database, a search that is currently prohibited.
The ACLU is opposing Ting’s bill, arguing that because it defines facial recognition as technology that monitors areas in real time, it sanctions the use of real-time community-wide facial recognition systems in certain circumstances including on a person suspected to have committed a felony. The organization says it won’t support anything short of an outright ban on the technology – something some members including Josh Lowenthal of Long Beach said they could see themselves supporting.
“Trying to regulate face surveillance is like trying to block a cannonball with cardboard,” ACLU attorney Carmen-Nicole Cox said in her testimony to the committee. “The harms of this technology fundamentally alter what it means to exist in public and already the status quo is that Black and brown communities are over-policed, over-arrested and over-abused by law enforcement. Face surveillance pours gasoline on that problem.”
To Williams, the Detroit man who was wrongfully arrested because of facial recognition, “the only answer is to not use this tech.”
“If the California legislature embraces the widespread use of face recognition technology, there will be many more cases like mine – innocent people put in harm’s way and damage that can never be undone,” Williams wrote. “But they may not be as lucky as I was to avoid a conviction, or worse, a fatal encounter with police.”