I guess they're applying the same test you would expect with libel; would the layman expect the speech to be truth or intentionally misrepresentative — i.e. wou

I guess they're applying the same test you would expect with libel; would the layman expect the speech to be truth or intentionally misrepresentative — i.e. would the image be construed as real. Otherwise, parody and criticism using AI would be limited.