OK, so there’s a technology that can be used to assist those who know what they are doing, or to pretend.

So it can be only used for things which don’t require responsibility. It can’t be used as a source in arguments, it can’t be used as a confirmation of a hypothesis, it can’t be used to make decisions.

So - the moment it’s possible fools start trying to stuff it in all these roles.

Why?

Easy, because they didn’t need it in any other roles! Only in those with burden of responsibility, because they want to make that burden smaller, get rid of responsibility so that they’d get the win, and somebody else the cost. Not only people too often want to use “AI” where they shouldn’t, they also very rarely use it where they should. Because they wanted a brain amplifier in the first place, something that thinks and takes responsibility instead of them! Not a tool to do something already easy.

There was the philosopher’s stone for medieval alchemists, and then there was romanticism for early XIX century’s liberals, and then there was misguided fascination with Plato’s “Republic” for late XIX activists, and then early XX century fascination with nihilism and monsters of steel as the solution to all problems.