Appeal to U.S. federal agencies: Get rid of Clearview AI

Appeal to U.S. federal agencies: Get rid of Clearview AI

Washington, 2/13/2022

Several Democratic lawmakers are calling on federal agencies to stop using the Clearview AI facial recognition program, saying it threatens people’s privacy and civil liberties. The lawmakers want the departments of Justice, Homeland Security, Defense, Interior and Health and Human Services to stop using facial recognition, calling Clearview AI’s technology „particularly dangerous.“

According to a GAO report, at least four federal agencies have used Clearview’s technology for „domestic law enforcement.“

Last week, Senators Ed Markey, Jeff Merkley, Pramila Jayapal and Ayanna Pressley, sent letters to the departments urging them to stop using Clearview AI and other similar products. In their letters, the lawmakers cite the threat the technology poses to marginalized communities, which goes beyond claims of inaccuracy.

Communities of color already face excessive police presence, so biometric surveillance would „disproportionately“ violate their privacy compared to non-minority populations, they write.

In a response, Clearview CEO Hoan Ton-That claimed that previous NIST tests showed the company’s facial recognition algorithm had no „demonstrable racial bias.“

The letters come after the IRS announced this week that it will no longer use third-party facial recognition software to verify taxpayer identities online.

Background:

Clearview AI recently won a contract to research augmented reality glasses for the U.S. Air Force Research Lab. According to reports, the technology would enable scanning and identification of individuals on military bases using glasses equipped with AI facial recognition.

The $50,000 contract simply tasks Clearview with researching the feasibility of the technology, which the military is exploring to protect its airfields.

Clearview CEO Hoan Ton-That said the technology is still in the development phase. He did not provide specifics on what it might involve, but said it would not be based on the company’s database of 10 billion images it collects from social media and the Internet. It would rely on limited data sets, such as images of missing children or suspects with outstanding warrants, he said.

A brief overview of the project provided by the Air Force said the glasses could „secure bases and flight lines.“ They would facilitate the use of weapons and help keep a distance from standoff situations or contagious diseases.

In 2020, The New York Times discovered code in the Clearview app that indicated it could be paired with augmented reality glasses. Ton-That confirmed that the company had a prototype at the time.