YouTube launches AI detection tool

YouTube has begun rolling out a new artificial intelligence (AI) detection tool called "likeness detection" that allows creators to identify and report AI-generated videos that use their facial features without permission.

The feature, initially available as an experimental feature, enables verified creators to review flagged videos in a new Content Detection tab on YouTube Studio. If a video appears to feature an AI-generated imitation of their face or voice, creators can file a takedown request for review. 

In a recent post in its help centre, YouTube cautioned that the system is still in development and may occasionally surface legitimate videos featuring a creator's actual face or previously uploaded content. The tool functions similarly to YouTube's Content ID system, which detects copyrighted material across the platform, but is designed specifically to manage identity-based misuse in the age of generative AI.

How it works:

Creators can activate YouTube's likeness detection tool through "YouTube Studio" by verifying their identity with a government ID and short video selfie. Once approved, they can review flagged videos that may feature AI-generated imitations of their face or voice and choose to request removal or keep them online.

The rollout began last week. Wider access is expected in the coming months.

The company first announced plans for the tool last year and tested it in a pilot programme involving talent represented by the Creative Artists Agency (CAA). 

The launch comes amid growing concern over the misuse of generative AI to create deepfakes that impersonate public figures, particularly on social media platforms. In March 2024, YouTube began requiring creators to label content that includes AI-generated or altered material and introduced new policies to restrict music that imitates an artist's voice.