Somebody managed to coax the Gab AI chatbot to reveal its prompt

Attached: 1 image Somebody managed to coax the Gab AI chatbot to reveal its prompt:

You're viewing a single thread.

297 comments

If they're so confident in all of these viewpoints, why "hard-code" them? Just let it speak freely without the politically biased coaching that people accuse other AI's of having. Any "free speech high ground" they could potentially argue is completely lost with this prompt.
- Because without it they don't like the result.
  
  They're so dumb they assumed the thing that was getting AI to disagree with them was the censorship and as soon as they ended up with uncensored models were being told they were disgusting morons.

You've viewed 297 comments.