This is a valid concern and while there are no guarantees that the AI won't hallucinate (hence the disclaimer in the book, especially for medical topics), I try to minimize it by pairing the writing with real-time research from Perplexity, so at least it is (or should be) based on verifiable information.