Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

An AI lawyer was set to argue in court – real lawyers shut it down

npr.org

Post reply on HN
Print view View original

511–520 of 708 posts

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#511

Post by lolinder »

mountainb wrote: post #464
lolinder wrote: post #417

Earlier quoted context omitted.

> The issue I think is that it's pulling from too many sources. There are plenty of sources that are pretty machine readable that will give it good answers. There's a lot of training that can be eked out from the legal databases that already exist that could make it a lot better. If it takes in too much information from too many sources, it tends to get garbled. No, this is a common misunderstanding about the way the…

If that's true, then it'll never get to the point that it would need to, and its reliability would always be too low to use.

Exactly. The paradigm is wrong, and in order to get past the problems we need a new paradigm, not incremental improvement on LLMs.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#512

Post by elicksaur »

gunshai wrote: post #331
elicksaur wrote: post #284

Good. Totally fine with trying to use AI to give legal advice, but it should be done with a lawyer’s license on the line. A company that explicitly disclaims being a law firm and states is not giving legal advice should also not get to tweet that they are “representing” someone in court. A good bar (pun intended) for the quality of the tech is if it is good enough that a licensed attorney trusts it to give legal advi…

No, bad. I may also disagree with some of the tactics DoNotPay used to represent themselves. But in a larger sense Lawyers cost money a lot of money. it's wonderful living in a time where the cost of filing suits is so low compared to defending against said suits. AI lawyers can help lower those barriers to entry. the court in question is fucking traffic court. please lower the barrier to entry for defense and allow…

Yeah, I agree with basically everything you’ve said, but the standard of AI legal advice should be at least the standard of services required by current attorney regulations. They should probably be higher even. It is certainly likely that in the future, technology can provide that level of quality at a massive, extremely cost effective scale, but ChatGPT and DoNotPay is not that.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#513

Post by dclowd9901 »

pocketarc wrote: post #20
epups wrote: post #12

Earlier quoted context omitted.

> a terribly dystopian idea in my opinion, as only the rich would then have access to actual lawyers The rich don't go to jail already. The crypto scammer paid a huge bail and got out on his private jet. That, to me, is far more dystopian than a cheap tool to help people appeal traffic tickets.

You're conflating a few different things. Being able to pay bail so you don't have to be in jail while you wait for your trial doesn't get you out of having a trial, and has nothing to do with needing lawyers. What the parent was referring to is the fact that if AI starts to consume the low-end (starting with traffic tickets), actual lawyers for trials will become even more expensive, and thus poorer people will actu…

Or maybe we only end up using lawyers when they're actually needed, and they become less costly for things like criminal trials. Think on the doctor whose routine cold and flu visits are replaced by an AI. Now they have a lot more time and bandwidth to handle patients who actually need physician care.

We can't just assume it's going to go the worst way. Neither outcome is particularly more likely, and the human element is by far the most unpredictable.

To wit: I was listening to a report yesterday on NPR about concierge primary care physicians. The MD they were interviewing was declining going that direction because they saw being a doctor as part duty and felt concierge medicine went against that.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#514

Post by giantg2 »

freejazz wrote: post #493
giantg2 wrote: post #486

Earlier quoted context omitted.

"It does" How? "Licensing has nothing to do with your credentials past having them." The way that most credentialing is used for employment.

That's separate from having your license. You don't need to be employed by a law firm to be a licensed atty

What's your point here? You still need a license.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#515

Post by legitster »

> Leah Wilson, the State Bar of California's executive director, told NPR that there has been a recent surge in poor-quality legal representation that has emerged to fill a void in affordable legal advice.

> "In 2023, we are seeing well-funded, unregulated providers rushing into the market for low-cost legal representation, raising questions again about whether and how these services should be regulated," Wilson said.

Got it. There are not enough affordable legal services in the US, and so the Bar's solution is to regulate them away.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#516

Post by retrac »

jamesdwilson wrote: post #490
lolinder wrote: post #417

Earlier quoted context omitted.

> The issue I think is that it's pulling from too many sources. There are plenty of sources that are pretty machine readable that will give it good answers. There's a lot of training that can be eked out from the legal databases that already exist that could make it a lot better. If it takes in too much information from too many sources, it tends to get garbled. No, this is a common misunderstanding about the way the…

If it is never pulling from a source, then why is it able to provide citations? check the chat bot on you.com: https://you.com/search?q=list%20top%205%20best%20selling%20c...

One useful way to think of language models is that they are statistical completion engines. It attempts to create a completion to the prompt, and then evaluates the likelihood, in a statistical sense, that the completion would follow the prompt, based on the patterns in the training data.

A citation in legalese is very common. A citation that is similar or identical to actual citations, in similar contexts, is therefore an excellent candidate for the completion. A fake citation that looks like a real citation is also a rather good candidate, and will sometimes squeak past the "is this real or fake?" metric used to evaluate generated potential responses.

This may seem like "pulling from a source" but there is no token, semantic information, or even any information in the model about where and when the citation was encountered. There is no identifiable structure or object (so far as anyone can tell anyway) in the model that is a token related to and containing the citation. It just learns to create fake citations so convincingly, that most of the time they're actually real citations.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#517

Post by gamblor956 »

freedomben wrote: post #498
fatbird wrote: post #488

Earlier quoted context omitted.

It had to be shot down harshly because there are some premises to a courtroom proceeding that aren't met by an AI as we currently have. One of those is that the lawyer arguing a case is properly credentialed and has been admitted to the bar, and is a professional subject to malpractice standards, who can be held responsible for their performance. An AI spitting out statistically likely responses can't be considered a…

> If the AI cites a non-existent precedent, there's literally no way to incorporate that error back into the AI because there's no factual underlying model against which to check the AI's output--unless you had an actual lawyer checking it, in which case, what's the point of the AI? IANAL, but I would bet the level of effort to fact check an AI's output would be orders of magnitude lower than researching and building…

You would lose that bet.

Fact-checking nonsensical output would take a lot longer than researching a single body of law, which you can generally do by just looking up a recent case on the matter. You don't need to check every cite; that will have been done for you by the lawyers and judges involved in that case.

But checking every cite in an AI's output: many of those citations won't exist, and for the ones that do, you'll need to closely read all of them to confirm that they say what the AI claims they say, or are even within the ballpark of what the AI claims they say.

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#518

Post by shmerl »

> As word got out, an uneasy buzz began to swirl among various state bar officials, according to Browder. He says angry letters began to pour in.

So an organization with the sole purpose of gatekeeping and anti-competitive market control does the anti-competitive market control. Why is this bar idea itself still legal?

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#519

Post by dan_mctree »

jamesdwilson wrote: post #490
lolinder wrote: post #417

Earlier quoted context omitted.

> The issue I think is that it's pulling from too many sources. There are plenty of sources that are pretty machine readable that will give it good answers. There's a lot of training that can be eked out from the legal databases that already exist that could make it a lot better. If it takes in too much information from too many sources, it tends to get garbled. No, this is a common misunderstanding about the way the…

If it is never pulling from a source, then why is it able to provide citations? check the chat bot on you.com: https://you.com/search?q=list%20top%205%20best%20selling%20c...

ChatGPT can provide correct citations because somewhere deep in its weights it does lossily encode real texts and citations to real texts. That makes real citations in some cases be its most confident guess for what is supposed to come next in the sentence. But when there isn't a real text it is confident about giving a citation about, but it still feels like a citation should be next in the output, it will happily invent realistic looking citations to texts that have never existed and it has never seen in any sources. On the outside, as readers, it's hard to tell when this occurs without getting an outside confirmation. I'm assuming though that to some degree it is itself aware that a linked citation doesn't refer to anything

Re: An AI lawyer was set to argue in court – real lawyers shut it down

#520

Post by lolinder »

bavila wrote: post #500
jchw wrote: post #478

Earlier quoted context omitted.

Pretty sure the whole reason why DoNotPay actually exists is because defending against parking tickets didn't actually require a strong defense. The tickets were flawed automation, and their formulaic nature justified and equally formulaic response, or something to that effect. Whether the LLM was actually going to output answers directly, or just be used to drive a behavior tree or something like that, is a question…

> Pretty sure the whole reason why DoNotPay actually exists is because defending against parking tickets didn't actually require a strong defense. The tickets were flawed automation... I have some past experience working in the courts in my state, and I know there are many judges who are perfectly fine with dismissing minor traffic infractions for no reason other than that they feel like it. If you've got an otherwis…

And I don't think anyone here has an issue with DoNotPay providing pre-trial advice and tips for someone defending themselves. It's bringing that into the courtroom that crosses a line from defending yourself to hiring an AI lawyer, and that line is where I'm very uncomfortable.
Post reply on HN