I wonder how long it takes AI to get good at law. Right now the verbal tasks it excels at are similar to the artistic ones: namely, solving problems with enormous solution spaces that are robust to small perturbations. That is, change a good picture of an angry tree man slightly and it's still probably a good picture of an angry tree man.
I've tried using a lot for writing motions. It can actually do a pretty decent job of writing motions, and it can come up with some arguments that you might not have thought of. You just have to ignore all its citations and look everything up yourself, otherwise this:
Isn't ChatGPT getting progressively better scores on medical and law exams? It will probably pass the USMLE and the bar one day. If it doesn't already.
Yes, but we should expect that, the answers are in its training data.
The problem is passing tests are an okay proxy for competence in humans, but if you think of LLMs as a giant library search engine, the thing it is competent at is identifying and regurgitating compiled phrases from its records.
Yes and that's amazing -- but law exams resemble programming exams. In the wild, both labors require you to keep a mountain of project-specific context in your head, something that tests like the LSAT cannot evaluate.