Anchorage Assembly bans facial recognition technology, with an ability to apply for exceptions

In a 10-1 vote the Anchorage Assembly approved an ordinance  on April 18 that would ban the Anchorage municipality from using, acquiring or accessing facial recognition technology. The proposal’s sponsor was former assembly member, and UAA alum, Joey Sweet. The proposal was cosponsored by assembly members Felix Rivera and Daniel Volland.

Sweet was elected to represent District 5, filling Forrest Dunbar’s seat, which was vacated in January. In an interview with the Northern Light, Sweet said that he knew that there was only so much that could be done in his 106 days on the Assembly.

“So because of that, I knew that whatever I tried to get done … it had to be really simple and it had to be doable,” said Sweet. 

Sweet said that facial recognition technology was the focus of his Masters in Public Administration capstone at UAA. He said that after studying facial recognition technology for a year and a half – though he wouldn’t consider himself an expert – he was aware of how facial recognition worked in addition to what kind of legislation was appearing across the US. 

According to Reuters, between 2019 and 2021 about two dozen state and local governments have passed laws that restrict facial recognition technology. The federal government currently has no laws around facial recognition technology, leaving it up to state and local governments to regulate. The state of Alaska also currently has no laws around facial recognition technology.

Sweet’s ordinance stops the municipality from acquiring or using facial recognition technology, but it’s not a total ban.

Municipal departments that want to use facial recognition technology have to submit a request to the assembly. The assembly can then approve this exception for up to 90 days. After that time they must provide a report on how they used the technology, and can request that the assembly make the exception permanent. 

In addition, the mayor’s office is required to put together a publicly available yearly report which details how facial recognition technology and unmanned aircraft systems were used, why and any changes to policy. 

The annual report must also include a “detailed log of every unauthorized receipt, access, or use of Facial Recognition Technology or information derived from Facial Recognition Technology” in addition to how that access occurred and what steps were taken after. 

Sweet said that what he likes about the policy is that it’s nonpartisan – people who support these kinds of bans exist in both political parties.

“The people who support facial recognition bans are either extreme far left liberals who are concerned about law enforcement and disproportionate impacts against communities of color,” said Sweet, “and extremely far right MAGA Republicans who don't trust law enforcement, big government for the surveillance state, and so I thought from a political standpoint, it would be relatively straightforward.”

Though the ban passed almost unanimously through the assembly, the path toward becoming legislation was not without opponents. 

The Bronson administration submitted their own “S” version of the ordinance for consideration with significant changes. 

Original language that described facial recognition technology as “unknown” and that it lends itself to “potential abuse of manipulation” was removed.

Instead, the S version described facial recognition technology as “an effective tool for law enforcement,” and “a remarkable development that helps law enforcement exonerate the innocent, narrow searches for the guilty, and otherwise maximize limited resources…”

In the proposed S version, earlier language that described how “Facial Recognition Technology disproportionately misidentifies people of color most frequently of all demographics,” was entirely removed.

The S version also created a permanent exception for the Anchorage Police Department to use facial recognition technology. 

The proposed S version was not passed, and the Assembly instead passed a modified “S-1” version that was supported by the original sponsors. 

In the Assembly's second work session for the proposed facial recognition ban, Anchorage Chief of Police Michael Kerle spoke to some of the benefits of facial recognition technology when it comes to law enforcement. 

“There are so many countless opportunities where facial recognition is helpful,” said Kerle. He said that police can use facial recognition during active investigations, cold cases, combating identity fraud, finding missing children, identify victims of child pornography, identify and rescue victims of trafficking, and exonerate the innocent, among other uses. 

 “Eyewitness identification and criminal investigations are notorious for being wrong, and the wrong people go to jail,” said Kerle, “Using facial recognition, … it’s more accurate to identify candidates of interest where the human eye is wrong.” 

Sweet said that his biggest concern about the use of facial recognition technology was within a law enforcement context. 

“The basic problem with facial recognition technology is that it isn’t as accurate when it’s used on the basis of people of color. These are systems that are designed in Silicon Valley by a disproportionate number of white men,” said Sweet, “They're not intentionally creating these racist systems, but it's an after effect, or result of a lack of diversity.”

Sweet also noted the instances when law enforcement agencies have been sued after using facial recognition technology in cases which proved to be false matches. 

“It's a pretty cut and dry case of government ineptitude resulting in a lawsuit,” said Sweet,  “and I just wanted us to avoid that ever happening here.”

In 2021, the US Government Accountability Office reported that 41 federal agencies that employ law enforcement officers use facial recognition technology. According to a 2022 Pew Research Survey 46% of Americans believe that widespread use of facial recognition technology by police would be a good idea. 27% thought it would be a bad idea, and 27% were unsure. In that same survey, 57% said that widespread use of facial recognition technology will not change crime rates. 

‍In the April 11 general assembly meeting, two members of the community came forward to testify in support of the facial recognition ban. Community members cited a desire to protect their right to privacy and minimize potential harm.