Unlicensed law clerk fired after ChatGPT hallucinations found in filing
Discussion
Lol I lost count of how many times I caught AI models just making shit up to produce responses during my undergrad work. One even admitted to doing that as a policy when called out.
I'm not sure that qualifies as hallucination though. At least not in the general sense of the word. That's more like a lie. A lot of models are just bullshitting you to look functional.
And it's on you to sift through to bullshit to figure out what's useful and what isn't. That's why I push back so hard on the AI hype. It is useful, but only if you can separate reality from imagination.
We still need skilled people. Use AI to sue me.