Comments by "bj0rn" (@TheLivirus) on "Researchers Find Multiple Ways To Bypass AI Chatbot Safety Rules" video.

  1. Yeah, regulation of chatbot output really seems like a losing battle. Even if it worked, that would also be an effective way to implement politically motivated censor or bias, which may be even more harmful.
    2
  2. People who really want to learn how to make a bomb will find out anyway. It's not that complicated, given that you learn a little bit of chemistry. At best this prevents children/teens from trying to make a bomb for the thrill.
    1
  3. As long as unrestricted AI is legal.
    1