Axon is banning facial recognition technology from the products it develops for law enforcement agencies because of “serious concerns” about the technology’s accuracy, the company said Thursday.
The announcement comes amid a heated battle over facial recognition software’s potential to amplify human biases.
Axon’s AI and Policing Technology Ethics Board — an independent oversight group of AI experts, computer scientists, civil liberties advocates, and public servants — recommended the company ban the technology. Axon agreed.
“No jurisdiction should adopt face recognition technology without going through open, transparent, democratic processes, with adequate opportunity for genuinely representative public analysis, input, and objection,” the board said in an announcement.
The controversy surrounding facial recognition has to do with communities that are already over-policed. Studies by MIT, the ACLU, and others have found that the software mis-identifies people of color and women more frequently than it does white men.
Amazon, a leading developer of facial recognition technology, disputes the findings of those studies. The company has come under fire for selling its Rekognition software to police departments.
“There have always been and will always be risks with new technology capabilities,” said Amazon’s artificial intelligence manager Matt Wood in a blog post addressing the concerns. “Each organization choosing to employ technology must act responsibly or risk legal penalties and public condemnation.”
Axon makes body cameras for police officers and other technology for law enforcement. The company has dual headquarters in Seattle and Scottsdale, Ariz.
As a neighbor of Amazon and Microsoft, another developer of facial recognition software, Axon could have leveraged its position to become a leading provider of the technology. But Axon says it isn’t sure if there will ever be an ethical way to provide facial recognition tools to police. The company says it will address that question once its concerns about bias and accuracy are addressed.
“At the least, face recognition technology should not be deployed until the technology performs with far greater accuracy and performs equally well across races, ethnicities, genders, and other identity groups,” Axon said.