Law enforcement routinely uses the Driver License Division’s database to identify criminals and prevent fraud, but Utahns have no guarantees about how their images are used.
A Senate committee gave unanimous approval Friday to a bill that would require the state to inform Utahns that their driver license photos — many characterized by poor lighting, awkward smiles or bad hair days — could be used for more than just a punchline.
But the bill doesn’t offer Utahns the option to opt out of the facial recognition technology — something the West Valley Republican says would render use of the database essentially meaningless.
“If I were going to go and get a fake ID with someone else’s name and Social Security number, of course I’m going to opt out, right?” he said in an interview. “So having an opt-out literally means that it does absolutely no good whatsoever.”
It also codifies the framework through which law enforcement agencies can tap into the database. Under the bill, they’re required to submit a written request for a facial recognition comparison that includes a statement of the specific crime they’re investigating and a “factual narrative” to support that there is a “fair probability” the person is connected with the crime.
The bill stipulates that a government employee can only comply with requests made for the purposes of investigating a felony, violent crime or a threat to human life; or to identify a person who is dead, incapacitated, at risk or otherwise unable to provide his or her identity to law enforcement.
Marina Lowe, legislative and policy counsel with the American Civil Liberties Union of Utah, told lawmakers Friday that the organization is supportive of the bill, while the libertarian-leaning Libertas Institute remained neutral. But both agreed that it doesn’t go far enough to address privacy and other issues posed by application of the technology.
“If this bill goes forward, by no means is the work done when it comes to facial recognition and the government’s use of this extremely powerful tool,” Lowe told members of the Senate Government Operations and Political Subdivisions Committee.
The ACLU sees the government’s use of private databases, which this bill does not address, as “the biggest threat” to privacy, she said.
“Any attempts by the state to get at the real problems around facial recognition technology would need to be much more expansive,” Lowe added. “And so if this bill passes, this is the first step and there should be a lot more work done to really address the root problems.”
Connor Boyack, president of the Libertas Institute, agreed that the bill leaves several key issues unresolved, including the government’s use of facial recognition with real-time CCTV feeds or video cameras in government buildings and body camera footage.
“There’s also now technology that is being used in Utah that does facial recognition based on social media photos, where an officer can take your photo with a cellphone, upload it to this tool and within seconds identify who you are based on your social media accounts,” he said. “Use cases like this are not contemplated in this bill.”
Thatcher acknowledged that there were many people who didn’t think his bill went far enough but said he wanted to get some kind of regulation on the books that everyone can agree with to start.
“I have no interest in carrying a controversial bill,” he said in an interview. “I’m going to carry the consensus bill. I’m going to carry the bill that does what everybody agrees is good policy.”
In addition to worries about the implications of widespread facial recognition technology on privacy, advocates have also raised concerns about the accuracy of these systems, as several studies have shown they can be prone to errors of misidentification especially with women and people of color.
Thatcher, in response to questions on that front from several lawmakers, noted that his bill was designed to address those concerns with a requirement that every match identified by a computer be verified with at least two officials trained on facial recognition.
“Two separate people, both with training, would have to look at the same [photo] and would have to reach the same conclusion before it would be considered a match,” he said.
Lowe, with the ACLU, was supportive of that requirement but noted that people aren’t infallible.
“Human beings can also make errors,” she said. “And so I don’t think that eliminates all of the problems we see in particular when looking at facial recognition problems when it comes to matching faces, particularly [of] people of color.”
The bill now heads to the full Senate for further consideration.