On Thursday, officials in the UK pledged to roll out a country-wide facial recognition system to help police track down criminals. The country’s ministers have launched a 10-week consultation to analyze the regulatory and privacy framework of their AI-powered surveillance panopticon — but one way or another, the all-seeing eye is on its way.
There’s just one tiny wrinkle: the AI facial recognition cameras have a tendency to misidentify non-white people.
New reporting by The Guardian notes that testing of the AI tech conducted by the National Physical Laboratory (NPL) found that it‘s “more likely to incorrectly include some demographic groups in its search results” — specifically Black and Asian people.
Though national police minister Sarah Jones has described facial recognition tech as the “biggest breakthrough for catching criminals since DNA matching,” lower-ranking police commissioners told The Guardian that the NPL finding “sheds light on a concerning inbuilt bias.” They likewise urged caution on the national rollout, which seems to be falling on deaf ears.
According to the NPL analysis, the national “retrospective facial recognition tool” — one of three types of facial recognition software used by national police — has a “false positive identification rate for white subjects (0.04 percent),” which is “lower than that for Asian subjects (4.0 percent) and black subjects (5.5 percent).”
“This has meant that in some circumstances it is more likely to incorrectly match Black and Asian people than their white counterparts,” the Association of Police and Crime Commissioners said in a statement to The Guardian. “The language is technical, but behind the detail it seems clear that technology has been deployed into operational policing without adequate safeguards in place.”
What this means for the nation-wide rollout remains to be seen. London is already one of the most heavily surveilled cities on Earth, with an estimated 1,552 cameras per square mile. Back in November, the Home Office offered funding to seven additional metro police forces to deploy fleets of facial recognition vans, joining police in London, South Wales, and Essex, which have been using the vans for some time.
Each of these vans is hooked into a police watchlist, and is outfitted with AI-powered facial recognition cameras on the roof.
As part of the 10-week consultation, the government collect feedback from citizens on whether police using facial recognition systems should be able to cross-reference their watchlist with other databases, like passport and driver’s license registries. Still, given that ministers had already pledged to dramatically expand the facial recognition technology, it’s unclear how highly public opinion will factor into the rollout.
If all goes according to plan for the system’s boosters, a new national database will be established, holding “millions of images” of innocent citizens, per The Guardian.
“The racial bias in these stats shows the damaging real-life impacts of letting police use facial recognition without proper safeguards in place,” said Charlie Whelton, a policy and campaigns officer with the advocacy group Liberty. “The government must halt the rapid rollout of facial recognition technology until these are in place to protect each of us and prioritize our rights — something we know the public wants.”
More on surveillance: AI Startup Says It Will End Crime by Blanketing the Entire United States in Ever-Watching Spy Cameras