YouTube expands AI deepfake detection tool

YouTube is expanding its AI-powered deepfake detection program to all users over the age of 18.

The feature uses a selfie-style facial scan that allows YouTube to monitor the platform for digital lookalikes of a user. If a match is found, the platform alerts the user, who can then request the removal of the content, CE Report quotes ANSA.

Until now, this AI-based technology had only been available to a limited group of well-known creators, politicians, government officials, journalists, and entertainment professionals. The program is now being expanded to anyone who wants to use it. The feature, called “content detection,” is available in the YouTube Studio menu.

The process begins with a scan of the user’s face, after which YouTube monitors uploaded videos for visual matches. When the algorithm identifies suspicious content using the user’s likeness, it sends a notification. The user can then review the report and decide whether to request the removal of the video.

The company’s guidelines state that every request is reviewed individually according to privacy rules, taking into account factors such as how realistic the video appears, whether it includes labels disclosing the use of AI, and whether the person can actually be identified. Satirical and parody content will continue to be protected.

The tool currently only detects facial similarity and does not cover other identifying features such as a person’s voice.

YouTube also clarified that there are no minimum popularity or subscriber requirements to access the tool, aside from being at least 18 years old. Users can also opt out of the program and request that YouTube delete their data.

PHOTO: CHAT GPT