In a world first, AI lawyer will help defend a real case in the US | Hacker News Reader

In a world first, AI lawyer will help defend a real case in the US

interestingengineering.com

22 pointsianbutler10 comments

> CEO Joshua Browder told New Scientist that it had found a court where listening via an earpiece was within the rules, even though it might not be in the spirit of the rules.

The thing about the legal system is that even when it allows hacks around the spirit of the rules (which it often doesn’t, because people that think they have found one are often reading the rules more narrowly than the courts themselves would when faced with a concrete instance), it can be quite quick at closing the door behind them once they are demonstrated.

“I found a hack around the way court rules are supposed to work and have built a business that relies on exploiting it” is...precarious.

That's awesome. Imagine having an AI acting as the judge and another one acting as the lawyer. It would basically automate our court system. But we do need our language models to have a larger short-term memory than merely 8k tokens.

As frustrating as the court system is, you do NOT want to automate law and order. Maybe it could work for small claims disputes, but imagine being sentenced to jail time without any amount of human compassion or sympathy.

>but imagine being sentenced to jail time without any amount of human compassion or sympathy.

This is not new

My experience with the legal system has been that judges and officers don’t really use discretion at all and hand down sentences in a similar way to an automated system. A judge will not be moved by or even listen to an impassioned speech by the defendant.

These AI's will emulate the behavior of our current system, so if such compassion and sympathy exist it will be preserved. Also, consider that the alternative is having a system where people are frequently bullied by people with deeper pockets to pay for good lawyers, which is an immense failure.

It is wild to me that any real technologist could hold this view. The more I learn about tech, the less I trust it, and the more I recognize the importance of strong societal regulation and interference.

Why? People would still be able to appeal to another human court if they disagree with the outcome. Although I believe most cases would be confirmed.

It's sort of happening since 2 year already in china internet courts (for small internet vendor dispute) in Beijing and Hangzhou.

In english: https://iacajournal.org/articles/10.36745/ijca.367

I don't think it'll take too long for "Prompt Generation and Processing" to be a paper as part of Law courses.

Reply on news.ycombinator.com