The two features nevertheless have a limitation, which the platform itself acknowledges. France Info specifies that the public tool makes it possible to check whether an image or a video was created with TikTok's or ByteDance's AI tools, but does not detect content generated by other AIs.
The Chinese platform is rolling out a likeness detector for public figures and a page open to the general public in order to identify videos made with AI.
The regulator cited several candidates in Creuse targeted by deepfakes. The mayor of Guéret filed a complaint.
For public figures, TikTok describes a process in several stages. The platform uses the example of a presidential candidate.
The candidate is notified when a video or an account uses their image. They can then check whether it is indeed a deepfake.
They can then report the content they consider misleading. The platform says it will remove it in accordance with"the community guidelines".
To examine a video, it must first be downloaded on TikTok. The option exists even without an account, provided the creator allows downloading. The file is then submitted to the detector.
The detector processes the file for a few seconds before displaying the result: content made with or without AI. TikTok assures that this information will not be retained.
YouTube had already introduced a likeness detector in September 2025, before the rollout announced by TikTok.