That's what they do! They are not called "Generative AI" for nothing. 'Generative' being just another word for "make things up". It's important to understand this, because once you do you understand that "casually making things up" is not a flaw at all. The real flaw is in our misunderstanding. And this misunderstanding leads to unrealistic expectations and misuse.
It should also be noted that this misunderstanding is perpetuated by the language being used by the AI community. "Hallucination", for example, falsely implies that the use case is legitimate, just a minor glitch here and there that we can fix with the next model.
Gen AI is an amazing bit of technology. It has some very interesting applications, so long as we understand what it does and so don't misuse it. Translating text, for example. Pattern matching in complex data. etc. But politicians and business executives are often turning to AI as a panacea when they themselves are out of ideas. And the AI moguls are encouraging them when they do so. We must call them out on this.
chrisjj•7mo ago
Academics at the University of Glasgow have said that because the models are not set up to solve problems, or to reason, but to predict the most plausible-sounding sentence based on the reams of data they have hoovered up, a better word for their factual hiccups is not “hallucinations” but “bullshit”.