A bipartisan group of U.S. lawmakers is considering federal legislation to limit the use of facial recognition technologies by law enforcement agencies.
“Facial recognition is a fascinating technology with huge potential,” Rep. Elijah Cummings, D-Md., said during a May 22 hearing of the House Oversight and Government Reform Committee. “But right now, it is virtually unregulated.”
There are an estimated 50 million surveillance cameras in use in the United States, according to the American Civil Liberties Union. Several cities are looking at ways to integrate facial recognition with their police departments’ surveillance cameras, including body cameras.
“The potential for mischief … is scary,” said Rep. Jim Jordan, R-Ohio. “That is a 1984, George Orwell kind of scenario.” He added that facial recognition could allow police or other government agencies to identify protesters or people attending a political rally.
Facial recognition technology continues to be plagued by misidentifications, especially when scanning the faces of women and African Americans, according to the testimony of Joy Buolamwini, founder of the Algorithmic Justice League. One program misclassified Oprah Winfrey as a man, she noted.
[Related: New York man sues Apple for $1B claiming face recognition technology led to wrongful arrest]
The privacy threat posed by facial recognition is unique, said Clare Garvie, a senior associate at the Georgetown University Law Center’s Center on Privacy and Technology. “Police can’t secretly fingerprint a crowd of people from across the street,” she said. “But they can scan their faces remotely and in secret.”
Garvie and Buolamwini both called for a moratorium on police use of facial recognition technologies until Congress passes regulations to restrict their use.
The House committee hearing came just days after the San Francisco Board of Supervisors voted to ban the use of facial recognition by police and other city agencies.
The same week, the Center on Privacy and Technology published two reports on police use of facial recognition, one on the scope of its use and the other on ways the technology can be abused.
Police should be forbidden from using facial recognition for generalized surveillance and monitoring when there’s no suspicion of a crime, argues Andrew Ferguson, a law professor at the University of the District of Columbia. He says police should need a warrant and a sworn declaration that privacy controls were used before being allowed to use the technology at all.
[Also read: ICE database raises privacy concerns]
Several House Oversight Committee members said they will work on legislation to limit uses of facial recognition. A second hearing will be held in June to hear testimony from law enforcement officials.
Away from the hearing, some technology experts questioned the wisdom of an outright ban.
Facial recognition technology can help identify criminals and find lost children or seniors, said Alan McQuinn, a senior policy analyst at the Information Technology and Innovation Foundation, a tech-focused think tank. Regulations are needed, he said, but not bans.
“It’s important to create guardrails around this technology, but guardrails are not roadblocks,” he said. “Police are deploying it effectively in many jurisdictions.”
Chad Steelberg, CEO and co-founder of artificial intelligence and facial recognition vendor Veritone, suggested that regulations would be a better approach than a ban or a moratorium. Using facial recognition software, one California police department was able to identify more than 150 suspects and save tens of thousands of dollars per case, he said.
Veritone markets its technology not as a general surveillance tool but as a way to identify criminal suspects after a crime has happened, Steelberg said. Law enforcement agencies should have court oversight when they’re using facial recognition, he added.
But banning the technology isn’t the right answer, Steelberg said. “San Francisco has really thrown the baby out with the bathwater.”
