Coalition pressures US govt agency to recommend ban on govt use of facial recognition

A number of advocacy groups in the US are pressuring the Privacy and Civil Liberties Oversight Board — an independent government agency that advises the presidential administration on privacy matters — to recommend that the federal government suspend use of facial recognition while accuracy and privacy issues are addressed.

Forty groups led by the Electronic Privacy Information Center signed a letter that called for the agency to stop government use of facial recognition “pending further review,” citing a recent New York Times report on a massive facial recognition database as one reason why the PCLOB should recommend the suspension to the Secretary of Homeland Security and President Trump. The report found that more than 600 law enforcement agencies across the U.S. were using a database of social media photos built by Clearview AI, a little-known startup.

“The PCLOB has a unique responsibility, set out in statute, to assess technologies and policies that impact the privacy of Americans after 9-11 and to make recommendations to the President and executive branch,” the coalition wrote. “The rapid and unregulated deployment of facial recognition poses a direct threat to ‘the precious liberties that are vital to our way of life.’”

In addition, the letter points to growing efforts on the local level — including in California and Massachusetts — to ban law enforcement use of facial recognition tech. Just this week, a bill was introduced in the New York state senate that would ban law enforcement agencies from accessing or installing biometric surveillance technology for their jobs.

Facial recognition algorithms in the U.S. have also struggled to accurately identify people of color, particularly black Americans. A 2019 report from the National Institute of Standards and Technology found that Native Americans had the highest rates of false positives, while African-American women were most likely to be misidentified in a law enforcement database.