Peaceful Hallucinations?

People worry about AI 'hallucinating' stuff and fabricating completely untrue information, yet we currently have a US president who is apparently hallucinating entire 'peace negotiations' with Iran.  Where are these negotiations taking place, through what channels and who is involved, (bearing in mind that the Trump administration keeps insisting that they've decimated Iran's leadership, it is fair to ask exactly who they are negotiating with)?   Nobody seems to know.  Iran, naturally, denies point blank that any such discussions have taken place.  So, is Trump simply imagining these 'negotiations'?  Is he in that stage of dementia where he's started having conversations with people who aren't there?  Now, there are those who will tell you that Trump's 'peace talk' is a calculated move to calm the markets sufficiently so that he and his cronies can manipulate and speculate upon oil prices.  But that would require him to be far more compos mentis than he actually is - he was never bright before, but is now in clear and rapid mental decline.  Could it be, though, that parties close to Trump are actually faking these negotiations, making the demented president think that he's involved in negotiations with Iran, in order for them to be able to manipulate the markets for their personal gain?  Maybe they just got a couple of random Iranian-sounding guys to speak to him on the phone, claiming that they were Ayatollahs, or something.  Because right now, as far as the US is concerned, nothing would surprise me.

But getting back to the AI, I've noted before the ridiculous levels of hysteria you'll find in some quarters with regard to it.  I still maintain that it is merely a tool which it is quite legitimate to use for doing some things, even in the 'creative' sphere.  Sure, there's plenty I wouldn't use it for: actually trying to create complete artistic works with it, taking medical advice, relationship advice, travel advice, in fact, advice of any kind.  The real problem with AI, in its current form, at least, is people.  Or rather the way in which people try to use it.  Rather than view it simply as a tool for achieving specific, very limited, tasks, they try to treat it as if it is actually, well, intelligent.  Which it isn't.  All it can do is, in essence, collate already existing information - it can offer no valid judgements on it, offer no real insights.  Yet people persist in trying to use it as if it does.  Lately, for instance, I've been reading about people who have used AI for legal advice, drafting legal letters, even.  Usually with predictably disastrous consequences.  Because the law is more than simply a set of rules - it is far more nuanced and open to interpretation.  Which is what lawyers are for, (well, decent ones, but this isn't the place for me to get into my opinions regarding many of those practicising in the legal profession).  Really, if you need legal advice, you are better off paying a competent lawyer for it that relying on an AI chatbot.  Likewise with reagard to medical advice - seek out a real live medical professional.  The fact is that people need to be educated as to how to use AI properly.  Because simply clutching your pearls or wringing your hands in despair at how it is allegedly stealing jobs etc, won't achieve anything.  Like climate change, denying it isn't an option - it exists, it is here and it isn't going away.  So, for God's sake, stop whining and learn how to use it properly.