LLMs hallucinate and are completely the wrong tool for this. People need to learn about the history of AI.

Expert-systems using techniques from the 80s are doing this right now [0].

[0] https://www.mathacademy.com/how-our-ai-works