2 ms·

I'd have no objections at all to them setting up a fake test case with a real judge or real prosecutors and doing controlled experiments where there's no actual

I'd have no objections at all to them setting up a fake test case with a real judge or real prosecutors and doing controlled experiments where there's no actual legal risk and where everyone knows it's not a real court case. You're right that it wouldn't be as attention-grabbing, but I suspect it would be a lot more useful for actually determining the AI's capabilities, with basically zero of the ethical downsides. I'd be fully in support of an experiment like that. Run it multiple times with multiple defendants, set up a control group that's receiving remote advice from actual lawyers, mask which group is which to the judges, then ask the judge(s) at the end to rank the cases and see which defendants did best. That would be a lot more work, but it would also be much higher quality data than what they were trying to do.

refactor_master 4y ago
> Run it multiple times with multiple defendants, set up a control group And also > That would be a lot more work, but it would also be much higher quality data I don’t know much about the field of law, but anecdotally it doesn’t strike me as particularly data driven. So I think, even before introducing any kind of AI, the above would be met with a healthy dose of gatekeeping. Like the whole sport of referencing prior rulings, based on opinions at a point in time doesn’t seem much different than anecdotes to me. But I’d love to be proven wrong though.
Forgeties79 4y ago
And in some ways it’s less work! The risks of using a real court case are massive if you ask me. We are a wildly litigious country. No amount of waivers will stop an angry American.