It does from time to time go a bit insane, and OpenAI has long been trustworthy in regards to the 'hallucinations' that ChatGPT can have, and the issues inherent in these LLMs. Be cautious of relying wholly on anything https://youtu.be/aGubCnA5Qyk