While there are countless interesting use cases for generative AI, the truth is that in the wrong hands it can cause a lot of damage. All it takes is a photo or a small audio clip for a cybercriminal (or someone with the resources to cause trouble) to clone our voice and upload a video of us saying words we didn’t say. This could be a potential problem, and YouTube has updated its policies to offer defenses to those potentially affected.

Violation of privacy. This is how YouTube treats someone who posts a video with synthetic content that looks like or resembles you. According to a recent update to the platform’s privacy policy, “If someone has used AI to modify or create synthetic content that looks like or resembles you, you can request that it be removed. To do so, the content must represent a version of your image that is realistic, altered, or synthetic.”

Although there are things to keep in mind. It won’t be as simple as reporting it, but YouTube will analyze the report and make a decision, taking into account a variety of factors. Those cited by the platform include whether the content is truly synthetic and flagged as such, whether the person can be “unambiguously” identified, whether the content is a parody or depicts a public or well-known figure, whether it is promoting some sort of criminal activity or advertising, whether it be a product or policy.

What is “absolutely identifiable”? There needs to be enough information in the video for others to be able to identify you. In that sense, YouTube believes that “the mere fact that you can be identified in the video does not mean that others can identify you” and that “for example, a name without additional context or a brief visual would not be considered identifiable in any way.” only”.

48 hours. The person who created the video needs to take action on this issue. The person who uploaded the video can delete it or blur or mute the parts that violate the person’s privacy. It is important to note that the person who files the complaint is the affected person and, with some exceptions (for example, if the affected person is a minor), they cannot do so on behalf of a third party.

“Just marked as synthetic”AI-generated content must be marked as such, but that doesn’t mean privacy policies still need to be adhered to.

Image | Pexels