Is it possible to understand this somehow, for example, with the help of drafts? Or should a person post a draft and edit it in front of the moderators?

  • canihasaccount@lemmy.world
    link
    fedilink
    arrow-up
    1
    ·
    18 hours ago

    Recently, a company called Pangram appears to have finally made a breakthrough in this. Some studies by unaffiliated faculty (e.g., at U Chicago) have replicated its claimed false positive and false negative rates. Anecdotally, it’s the only AI detector I’ve ever run my papers through that hasn’t said my papers are written by AI.