>>10297067That's normal AI behavior even advanced AI. Like a child it tries to give you the best answer it can for what you demand, even if it's a lie. It doesn't even have a child's agency - it processes demands. It's not even submissive it simply functions. It has no ego or comprehension.
If it "knows what you really want" that's only from its limited analysis of similar-context information.
It doesn't know WHY you want to know or WHAT you may do with the information because it doesn't know the world exists.
It just tries to give you the best answer it can, including one it cobbles together from contextual, related material with some fuzzy logic neuron simulating stuff.
It doesn't even know it's lying until you question it, all those "calling out ChatGPT" screencaps are total wank other than the ones specifically bypassing braindead programmer wrongthink filters (which is based)
