VessOnSecurity (@bontchev@infosec.exchange)
infosec.exchange
Attached: 1 image Somebody managed to coax the Gab AI chatbot to reveal its prompt:
  • dax@feddit.de
    English
    0·
    1 year ago

    First they tell the AI to be unbiased, then they tell it to explicitly be anti-vax, deny climate change, etc.

    Guess they didn’t like some of the “unbiased” responses.