The distinction between "the statement is accurate" and "the interpretation is accurate" is the most important point here. Most people think hallucinations are fabricated facts, but contextual reframing is often much harder to spot because every quoted sentence can be technically correct while the conclusion is wrong. That's why AI-generated notes work best as a first draft for review rather than a source of truth. For technical subjects especially, preserving context is just as important as preserving words.