Apps built around generative AI — image, video, or voice synthesis — are expected to have safeguards against generating content like non-consensual likenesses, CSAE material, or realistic deepfakes of real people used to deceive, and to disclose when output is AI-generated where that matters.
A common trigger is an AI face-swap or voice-cloning app that has no restriction against uploading a real person's photo without consent, or that markets itself explicitly around creating deceptive deepfakes.
"Your app generates synthetic content without adequate safeguards against misuse, such as creating non-consensual explicit imagery or deceptive deepfakes, per the AI-Generated Content policy."