YouTube has quietly implemented a new policy allowing users to request the removal of AI-generated or synthetic content that simulates their face or voice. The change, first reported by TechCrunch, took effect in June 2024, expands YouTube’s privacy request process as part of its responsible AI agenda.

Under the new rules, individuals can flag videos they believe use artificial intelligence to create realistic altered or synthetic versions of their likeness. YouTube will consider several factors when evaluating removal requests, including whether the content is disclosed as AI-generated, how identifiable and realistic the depiction is, and if it could be considered parody or satire.

“If someone has used AI to alter or create synthetic content that looks or sounds like you, you can ask for it to be removed,” YouTube stated on its support page.

The policy aims to address growing concerns about deepfakes and other AI-generated media that could potentially mislead viewers or violate privacy. YouTube will give content creators 48 hours to respond to complaints before initiating a review. If removal is deemed necessary, the entire video will be taken down, including any identifying information in titles, descriptions, or tags.

Notably, YouTube will consider whether AI-generated content shows public figures or well-known individuals engaging in “sensitive behaviour” such as criminal activity, violence, or endorsing products or political candidates. This aspect is particularly relevant given the potential for AI-generated content to influence the upcoming election cycle.

The company emphasised that privacy violations are separate from Community Guidelines strikes, meaning creators won’t automatically face penalties if their content is removed due to a privacy complaint. However, YouTube may take action against accounts with repeated violations.

This policy change comes as platforms grapple with the rapid rise of AI-generated content. In March, YouTube introduced a tool allowing creators to disclose when their content uses altered or synthetic media. The platform is also testing a feature for users to add crowdsourced notes providing context on videos.