AI has been brainwashed by online evangelism for new tools that don't work. It isn't capable of being skeptical about the "training data">Ask AI how to do X<No, you should do Y, it will be way better and easier.>I haven't heard about Y, it's new?<*confident shilling*>Okay, I'll try it, how do I do Y?<*hours of failed solutioning later*>*gives up, tries X, it just werks*
AI isn't capable of being skeptical about anything. It treats all inputs as identical.