These wrongfully arrested black men say a bill in California would allow police to abuse facial recognition

In 2019 and 2020, three black men were charged and incarcerated for crimes they did not commit after police misidentified them using facial recognition. Their wrongful arrest lawsuits are still pending, but their cases demonstrate how AI-powered tools can lead to civil rights violations and lasting consequences for the defendants’ families.

Now all three men are speaking out against a bill underway in California that would prohibit police from using facial recognition technology as the sole reason for a search or arrest. Instead, corroborating evidence would be required. The problem, critics say, is that a possible facial recognition “match” is not evidence — and that it can mislead investigations even when police are looking for corroborating evidence.

The State Assembly passed House Bill 1814 last month by a vote of 70 to 0. Today it faced a contentious hearing in the Senate Public Safety Committee.

Such a law “would not have stopped police from falsely arresting me in front of my wife and daughters,” Robert Williams said in a statement to CalMatters. In 2020, Detroit police charged Williams with stealing thousands of dollars worth of watches — the first known case of a false arrest using facial recognition in the United States — after facial recognition matched surveillance video to a photo of Williams in a government database. Investigators placed his photo in a “six-pack lineup” with five others, from which a security guard who had seen a surveillance image and not the theft itself chose him.

“In my case, as in others, the police did exactly what AB 1814 required them to do, but it didn’t help,” said Williams, who is black. “When the facial recognition software told them I was the suspect, it poisoned the investigation. This technology is racially biased and unreliable and should be banned.”

“I implore California lawmakers not to settle for half measures that do not truly protect people like me.”

The first facial recognition searches in the United States occurred more than two decades ago, a process that begins with a photo of a suspect, usually taken from security camera footage. The facial recognition on your iPhone is trained to match your photo, but the facial recognition used by law enforcement searches databases of mugshots or driver’s license photos, which can contain millions of photos, and can fail in numerous ways. Tests by researchers have shown that the technology is less accurate when trying to identify people with dark skin, people of Asian or American descent, or people who identify as transgender, when an examination image of a suspect is of poor quality, or when the image in a database is out of date.

After a computer compiles a list of possible matches from an image database, police select a suspect from a pool of candidates and then show that photo to an eyewitness. Although people often think they’re good at this, eyewitness testimony is one of the leading causes of wrongful convictions in the United States.

Because prosecutors use facial recognition to identify potential suspects but ultimately rely on eyewitness accounts, the technology can play a role in criminal investigations but remains hidden from the accused and defense attorneys.

Instructions not to consider a possible match with a facial recognition system as the sole reason for an arrest sometimes make no difference. For example, it did not do so in the case of Alonzo Sawyer, a man who was wrongfully arrested near Baltimore and spent nine days in jail.

Njeer Parks, who spent nearly a year battling allegations that he stole items from a New Jersey hotel gift shop and then nearly hit a police officer with a stolen car, spoke out against the California bill in a video posted to Instagram last week, saying police “won’t do their job if the AI ​​already says, ‘It’s him.’ That’s what happened to me.”

“I was lucky,” he told CalMatters in a phone interview about a receipt that exonerated him and kept him out of jail. “I don’t want anyone to go to jail for something they didn’t do.”

The lawyer for Michael Oliver, a third black man falsely accused of attacking a Detroit high school teacher in 2020, is scheduled to testify at today’s hearing before lawmakers in Sacramento, the American Civil Liberties Union said.

Supporters of the bill include the California Faculty Association and the League of California Cities. The California Police Chief Association argues that facial recognition can reduce crime and provide police with actionable leads. It also says the technology is important because California wants to host international events such as the 2026 FIFA World Cup and the 2028 Summer Olympics in Los Angeles.

“Across the country, real-world examples of law enforcement using (facial recognition technology) to solve serious crimes demonstrate how important this new technology can be in protecting our communities,” argued the California Police Chiefs Association, citing cases in which facial recognition played a role in identifying the culprits, including a shooting at newspaper headquarters in Maryland and a rape in New York.

Facial recognition alone should never lead to false arrests, Jake Parker of the Security Industry Association told members of the California Assembly a few weeks ago. That’s why AB 1814 is intended to support investigative approaches with evidence, not just a possible match in facial recognition.

“There is clearly a need to increase public confidence that this technology will be used properly, lawfully and effectively, in a way that is limited and non-discriminatory, so that our community benefits,” he said. “We believe AB 1814 will help increase that confidence, and that is why we urge you to support this bill in its current form.”

But more than 50 advocacy groups – including the ACLU, Access Reproductive Justice and the Electronic Frontier Foundation – signed a letter opposing the bill last week, calling facial recognition unreliable, a proven threat to black men and a potential danger to protesters, abortion seekers, immigrants and LGBTQ communities.

“By allowing police to scan and identify people without restriction, AB 1814 will also result in unnecessary police calls that all too often have the potential to escalate into deadly encounters. This will continue regardless of how accurate facial recognition technology becomes,” the organizations said in a letter. “There is no way for people to know if facial recognition is being used against them and no mechanism to ensure police are complying with the law.”

The bill’s author, Democratic Rep. Phil Ting, also authored a 2019 bill that initially permanently banned police from using bodycam footage with facial recognition. That was changed to a temporary ban that would expire in January 2023.

Ting told CalMatters that he is uncomfortable with the fact that California currently has no restrictions on the use of facial recognition by law enforcement.

In a statement, he said his bill “simply requires officers to have additional evidence before proceeding with a search, arrest, or warrant affidavit. I believe a precautionary measure can help protect people’s privacy and due process rights while still allowing local governments to go further and enforce their own facial recognition bans.”

Ting’s city of San Francisco became the first major city in the country to ban facial recognition in 2019. But an analysis by city attorney David Chiu found that the city’s passage of Proposition E in March allows police to conduct facial recognition searches on images captured by cameras and drones. The Washington Post reported last month that San Francisco police are getting around the restrictions by asking law enforcement in neighboring cities to conduct the searches for them.

Re-elected San Francisco District Attorney Chesa Boudin says false arrests related to facial recognition have almost certainly occurred in California, but they remain unknown to the public unless prosecutors file charges and the defendants are later tried in a civil suit seeking damages. Often such cases are settled out of court.

“We absolutely need a legal and regulatory framework for these technologies, but I do not believe AB 1814 is sufficient to protect civil liberties or provide meaningful safeguards for the use of these new and powerful technologies,” said Boudin, who now directs UC Berkeley’s Criminal Law & Justice Center.

Senate Public Safety Committee staff have proposed amending the bill to provide that judges should not grant warrant requests from law enforcement based solely on a potential match using facial recognition technology.

Lawmakers have until the end of the legislative session in August to decide whether to pass AB 1814.