seems to completely ignore that lack of regulation on AI watermarks (outside of China). All future datasets will be polluted with AI output, worsening accuracy. The more “AI” (large language models) grow as a field, the worse the problem will get

Relatable