For every fact you learn there is a necessary context and numerous exceptions when it's applied in other contexts. If you expect AI to spit all relevant facts at you every time you're clueless, you will have no idea when AI is wrong or lying. You will have to trust everything AI (or its owner) tells you.