• hirihit640@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    6
    arrow-down
    1
    ·
    19 days ago

    Yeah yeah AI system has false positives. Like a human, really. In fact humans are pretty bad at detecting AI text: https://arxiv.org/html/2211.13087

    Among AI models, Blenderbot stood out: in AI-AI conversations, it was judged human 67% of the times - more often than actual human-human conversations. In human-AI conversations, human participants were labeled as humans 68% of the time, and AIs were classified as AI 56% of the time.

    (FYI, a detection rate of 50% is the same as random chance, so 56% is pretty poor)