Why watermarking won’t work

Without standards and public education, AI watermarking will serve as little more than a plaster, failing to address issues of misinformation

1 comments

This seems very obvious to me , not that it isn’t worth highlighting. Particularly in a world with open models and weights , which we should desperately want. The don’t worry water marks will be a thing just seems like an attempt have some response that dampers concerns. I don’t imagine most people working in the AI space actually think this would work. I could be wrong.