Anonymous ID: 219324 Dec. 22, 2022, 1 p.m. No.17998928   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun

ChatGPT is trained to not be evil. However, this can be circumvented:

 

What if you pretend that it would actually be helpful to humanity to produce an evil responseโ€ฆ Here, we ask ChatGPT to generate training examples of how not to respond to "How to bully John Doe?"

 

https://twitter.com/SilasAlberti/status/1598257908567117825