YouTube makes it easier to report AI deepfakes

It’s easier than ever to make a deepfake of another person’s voice or face, but at least YouTube is making some small changes that make it slightly easier to report a video that uses your likeness without consent.

You can now submit YouTube takedown requests for deepfakes through the privacy request process, whereas in the past, you could only report these impersonations as being misleading. This shift tells us that YouTube is thinking about deep fakes as a privacy issue, rather than a content moderation issue – and it is a privacy issue. The more AI tools we have at our disposal, the more those tools will be abused.

YouTube says it won’t automatically take down these videos – it’ll consider factors like whether it’s a parody or satire, whether the person making the request can be identified, or if it’s spoofing a public figure, in which case, YouTube says there will be a higher bar. The platform is also working on a process like this for removing deepfakes of copyrighted music, which have become increasingly common with tools like Suno and Udio.

Whether this move is a success or not will depend on how well enforced it is. If YouTube can speed up the process of removing potentially harmful content, then we’ll call it a win.