AI poses a risk to our understanding in general – or rather how much we value understanding. If people get too used to having an answer to their questions immediately (by asking AI), will they start spending less time trying to understand those answers? We know smth when we believe smth is true based on a reliable source, but we understand smth when we know why or how its true, and that the source is reliable. So the argument is that even if we assume that AI is a reliable source, and allows us to know smth bc we believe it is true based on AI, we still don’t understand that thing bc we don’t know why or how it is true, and why or how AI is reliable. AI will lead us to be passive in our knowledge – I don’t need to understand why something is true, because I asked AI, and that’s enough.