minus-squareCoolkidbozzy [he/him]@hexbear.nettoTechnology@lemmy.ml•*Permanently Deleted*linkfedilinkEnglisharrow-up5arrow-down1·1 year agoseems 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 linkfedilink
minus-squareCoolkidbozzy [he/him]@hexbear.nettoTechnology@lemmy.ml•ChatGPT gets code questions wrong 52% of the timelinkfedilinkEnglisharrow-up19·1 year agoRelatable linkfedilink
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