I remember taking college classes where one had an online test. I typed in the correct answer, however because I formatted it differently than what the test exp

I remember taking college classes where one had an online test. I typed in the correct answer, however because I formatted it differently than what the test expected...I failed the question. How many times have we seen these chat AIs give wrong answers? I could see a student given a correct answer but the AI failing it because AI really isn't that smart. I really expect there to be major major issues with the advent of AI, not in that it's taking over people's work (although thats a big issue) but simply that people will rely on AI to be correct when it's actually wrong. AI is based on existing knowledge (and since it's been taught off the internet, it most likely has a lot of wrong information). We need people to a) check the work of the AI b) still need people to do new research and collect new data.