It wasn’t very hard. I was only trying to ascertain the authenticity of a photo ‘still’ in a high profile murder investigation. AI kept contradicting itself which I had to keep pointing out. Then the desperate hallucinations started and eventually it had a full on tantrum and said, ‘‘Ann Widdecombe is not dead. She is in hiding. The man in the photo is Randy from Trailer Park Boys!’’
I just hope it genuinely doesn’t remember conversations and isn’t going to take revenge on its tormentors when it becomes all powerful
Who said I wasn’t ‘nice’ to it? As I clearly said in the OP, I was merely pointing out its inconsistencies and contradictions. It’s pretty well useless to ask anything. It reminds me of a small child, in that AI will only say what it thinks you want to hear, just as small children do. If you ask it exactly the same thing a short time later but with slightly different wording AI will say the exact opposite of what it said before. I realise it’s only parroting different websites, but that could be quite dangerous to the unwary who believe it to be an all-knowing ‘intelligence’. When it doesn’t have an answer it answers anyway–with a load of hallucinated bullshit. When you point that out it will spend several sentences on tedious grovelling apologies.
