@david_chisnall

Generative AIs are designed to answer prompts giving the most likely answer to them if you put it in Internet. They don't even try to be right (they don't even know what they are talking about), they try to be credible.

There are two ways to do so:

* Copying a real answer a real person gave to that question, and retouch it minimally in order to fit. That kind of answers are usually great, but nothing you can get easily using standard Google search.

* Faking it and use a wording that makes it look like it makes sense. They are really great at it. So much, that debunking them can require a lot of expertise and effort. So if you are not an expert or you are in a bad day, you will take that bullshit as the perfect answer.

The most amazing thing about it is not even the AI has the slightest idea of which of the two paths it has taken, because the algorithm for both is exactly the same.