Comment by written-beyond

It's still okay to do that, LLMs models are literally terabytes big I'm sure they've crammed enough information in there and enough verification tools to be extremely accurate. Now whether the information matches your true internal intent, rather than the prompt, is another thing.