YouTube introduces AI 'Likeness Detection' Tool to Combat Deepfakes

Staff Reporter

Staff Reporter

26 October 2025, 22:49

YouTube introduces AI 'Likeness Detection' Tool to Combat Deepfakes
[photo collected]

YouTube has begun deploying a new artificial intelligence tool—dubbed "likeness detection"—that empowers verified creators to identify and request the removal of videos featuring AI-generated imitations of their face or voice. 

The move is the latest step by the platform to tackle the growing threat posed by sophisticated digital impersonation.

The feature is initially being launched as an experimental tool and is accessible to eligible creators through a dedicated Content Detection tab within the YouTube Studio dashboard.

How the System Works
Creators who opt into the new system must first verify their identity using a government ID and a short video selfie. Once approved, they gain the ability to review videos flagged by the AI for potentially containing an altered or synthetically generated version of their likeness.

From this review interface, creators have the option to file a takedown request under YouTube's privacy policies or choose to keep the content online. The technology is designed to function similarly to YouTube's long-established Content ID system, which automatically scans for copyrighted material, but is focused specifically on managing identity-based misuse.

Context and Previous Measures
The rollout follows an initial pilot programme tested last year with high-profile talent represented by the Creative Artists Agency (CAA). YouTube, owned by Google, is moving to provide creators with a critical defense mechanism against the misuse of generative AI to create deepfakes that could spread misinformation, harm reputations, or create fake endorsements.

The company cautioned in a recent Help Centre post that the detection system is still in development and may occasionally flag legitimate videos featuring the creator's actual face or previously uploaded content.

This measure is an escalation of YouTube's existing AI governance policies. In March 2024, the platform began requiring creators to clearly label all content that includes realistic AI-generated or altered material. 

Furthermore, YouTube has also introduced policies to specifically restrict music that imitates an artist's unique vocal style without authorisation. Wider access to the new likeness detection tool is expected to expand over the coming months.