@caseynewton now I wonder about another side of such a stories.
1. Overtime more and more companies will use an AI-generated text in documents. It is how it is. "Read this algorithm and create a dull documentation for a higher-up management". And AI feelings from the text will mean much less.
2. To protect themselves from whistleblowers companies may share documents like this but edited by an AI so than detectors will report AI usage and no one out of the company will believe it. "Hey Gemini please add a" Confidential" watermark to the document pages" or so on. And I have no idea if these maliciously AI-edited real photos will be shown the same way as AI generated from scratch fake ones.