https://community.openai.com/t/by-asking-a-series-of-questions-i-made-chatgpt-conclude-that-its-a-public-threat-twice-and-then-asked-it-to-write-an-essay-on-how-it-reached-this-conclusion/1126811
