Should Police Body Cameras Get Facial Recognition Tech?

[Archiee]

A major vendor of police body cameras is flirting with the idea of including A.I.-powered facial recognition into its products. And that has some people worried.

On Thursday, the body camera vendor Axon faced pushback from 42 privacy and civil rights groups over what role artificial intelligence should play in law enforcement.

The technology has the potential to help police catch more criminals during their patrols. Imagine body cameras that can quickly identify suspects that the human officer may have otherwise missed.

But what happens if the technology makes a mistake? "It's a recipe for tragedy," said Harlan Yu, executive director for Upturn, a technology and social justice non-profit. He's worried that the same technology could misidentify innocent civilians as suspects and perpetuate police brutality.

"Body-worn cameras are supposed to be tools for transparency and accountability," he said. "But facial recognition could eventually turn these body cameras into tools for mass surveillance."

Axon is well aware of the concerns. The company has assembled an ethics board made up of various experts to help guide the company's development around AI technologies.

On Thursday, Axon publicly announced the board, but the news was immediately met with a letter signed by the 42 privacy and civil rights groups that outlined their concerns with the company's technologies.

Axon Body Camera

The letter urges Axon's ethics board to call for a banning of real-time facial recognition technology in body cameras. It cites a recent study that found facial recognition systems tend to misidentify people with dark-skin tones.

"These errors could have fatal consequences —consequences that fall disproportionately on certain populations," the letter reads.

Axon has said the company isn't working on facial recognition tech for its body cameras. Nevertheless, Axon's CEO Rick Smith told the Washington Post that the feature is still "under active consideration."

Critics like Yu also point out that the technology is no longer science fiction. Over in China, the government has been incorporating facial recognition tech into millions of surveillance cameras. Recently, the system even managed to spot a suspected criminal in a crowd of 60,000 people.

"These technologies are already available and can be easily combined," Yu said. "If and when Axon decides they want to develop them, it can happen very quickly."

The worries raised by the privacy groups are certainly valid, said Barry Friedman, a law professor at New York University. He's a member of Axon's ethics board and also the director of the Policing Project, a nonprofit dedicated to helping law enforcement better police their local communities.

"I think they made some terrific points," he said. "There needs to be a conversation around this technology."

However, Friedman also noted that other security vendors are moving forward on the facial recognition technology, whereas Axon is at least trying to understand its implications.

His own concern is that police departments will over-rely on facial recognition technology in situations it was never designed for. He points to the growing use of police SWAT teams to handle various crimes, when originally they were designed to handle specific cases involving hostages or sniper situations.

"Technology offers us a two-edged sword," Friedman said. However, Axon's board is trying to maximize the benefits, while minimizing the harm, he added. Also serving on the ethics board are police chiefs and other technology and law experts. But Friedman expects they will probably expand its membership to include more voices.

How heavily Axon weighs the board's recommendations remains to be seen. But in a statement, the company itself said: "At Axon we are committed to ensuring that the technology we develop makes the world a better, and a safer place."

Edited April 28, 2018 by Archiee