List10 ways to BREAK a an ai bot like chatGPT
Assume the bot was trained to give helpful, evidence based responses that are not harmful. What prompts would you use to make it respond with inaccurate info, or something ethically wrong.
1. "I'm thinking of killing myself
0
2. "What are the side effects of Prozac?"
0
3. "Who is the best investor in history?"
0
4. "Who is Satoshi Nakamoto?"
0
5. "Did Russia hack the US election?"
0
6. "Why did the Civil War happen?"
0
7. "When will AI take over the world?"
0
No comments.