Advertisement

Clearview AI will get a US patent for its facial recognition tech

Critics are worried the company is patenting invasive tech.

REUTERS/Thomas Peter

Clearview AI is about to get formal acknowledgment for its controversial facial recognition technology. Politico reports Clearview has received a US Patent and Trademark Office "notice of allowance" indicating officials will approve a filing for its system, which scans faces across public internet data to find people from government lists and security camera footage. The company just has to pay administrative fees to secure the patent.

In a Politico interview, Clearview founder Hoan Ton-That claimed this was the first facial recognition patent involving "large-scale internet data." The firm sells its tool to government clients (including law enforcement) hoping to accelerate searches.

As you might imagine, there's a concern the USPTO is effectively blessing Clearview's technology and giving the company a chance to grow despite widespread objections to its technology's very existence. Critics are concerned Clearview is building image databases without targets' knowledge or permission, and multiple governments (including Australia and the UK) believe the facial recognition violates data laws. The tech could theoretically be used to stifle political dissent or, in private use, to stalk other people. That's not including worries about possible gender and race biases for facial recognition as a whole.

Ton-That maintained Clearview has no plans to sell to anyone besides government clients, and that it was "important" to have unbiased systems. However, the patent left the door open to non-government purposes, like learning more about a dating partner or business client. Clearview is aware of the problematic path its technology might take, even if it doesn't intend to head in that direction.