YouTube has decided to expand its deepfake detection tool, known as similarity detection, to all adult users. This feature allows scanning the platform to locate videos that use a person's face without their consent. Upon detecting a match, the system alerts the user, who can request the removal of the content. The final decision is based on privacy policies and whether the material is realistic or generated by artificial intelligence.
How YouTube's facial scanner works 🧐
The tool focuses exclusively on the facial image, leaving aside voice or any other element. When a user reports a video, YouTube's team evaluates whether the content infringes privacy rules or if it is a parody or satire, cases that are excluded from the process. The system does not analyze the audio or the full context of the scene, only the representation of the face. This limits its scope, but offers a direct route for those seeking to protect their visual identity on the platform without resorting to complex legal processes.
Goodbye to videos where you look like Brad Pitt 😂
Finally, YouTube cares about your face, though not so much about your voice. Now you can report that video where a YouTuber put Brad Pitt's face on you dancing in a karaoke contest, just when you were in your pajamas. The tool promises to be useful, although it will surely generate more than one false report from someone who doesn't want it known that their face appeared in a tutorial on how to make homemade bread. The irony: the AI that creates deepfakes is now fought by another AI, in a duel of algorithms that only we humans enjoy watching.