YouTube AI detection tool for deepfakes

Photo Credit: Eyestetix Studio

YouTube is rolling out its AI detection tool for eligible creators to request the removal of AI-generated content featuring their likeness.

On Tuesday, YouTube announced the official rollout of its likeness-detection tool for eligible creators in the YouTube Partner Program after a test phase. The technology allows creators to request the removal of AI-generated content that features their likeness.

YouTube’s detection technology identifies and manages AI-generated content featuring the likeness of creators, including their face and voice. The tool is designed to prevent the misuse of someone’s likeness, be it to endorse products and services they have not agreed to support, or for spreading misinformation.

Eligible creators will have received an email this week, YouTube reports, but the company also provided instructions for the technology’s use on its Creator Insider channel. Those who have access will find the tool as an option in YouTube Studio, hanging out in the existing “Content Detection” menu.

Creators will first need to go to the “Likeness” tab, consent to data processing, and use their smartphone to scan a QR code that will direct them to a web page to verify their identity. The process also requires a photo ID and a brief “selfie video.”

Once YouTube grants access, creators can then view all detected videos and either submit a removal request according to YouTube’s privacy guidelines, or make a copyright request. An option to archive the video also exists.

Creators can also opt out of using the technology at any time, and YouTube will stop scanning for videos 24 hours afterward.

YouTube’s likeness-detection technology has been in pilot mode since early this year. The company first announced last year that it had partnered with Creative Artists Agency (CAA) to help celebrities and creators identify content on the platform that uses their AI-generated likeness.

The company also expressed its support in April for the legislation known as the NO FAKES Act, which aims to address the issue of AI-generated replicas imitating a person’s image or voice to deceive others or generate harmful content.