A multimillion-dollar conspiracy trial that stretched across the worlds of politics and entertainment is now touching on the tech world with arguments that a defense attorney for a Fugees rapper bungled closing arguments by using an artificial intelligence program.
I saw an article from ars that tracked the AI company down, it’s registered to the same office as the lawyer, and immediately started advertising this case bragging about it being used in an actual trial, no mention of how much it fucked up and the client was guilty.
He’s got a pretty good shot at this, and the lawyer should 100% face consequences. Even if he just used it, but especially if he owns the AI company he used. Doubly so for not disclosing the connection or informing the client it was being used.
Removed by mod
That’s irrelevant. The AI is not licensed to practice law; so if the lawyer didn’t perform any work to check the AI output, then then the AI was the one defending the client and the lawyer was just a mouthpiece for the AI.
Yeah I feel like this is the same as if the lawyer had used a crystal ball to decide how to handle a case. If he lied to clients about it or was also selling crystal ball reading services that seems pretty bad.
Removed by mod
If you’re found guilty because of a typo you’re probably going to have a successful appeal.
This could very well be what he has to prove. That the lawyer didn’t do his due diligence and just trusted the ai.
No, that’s a bad question. Autocorrect takes your source knowledge and information as input and makes minor corrections to spelling and suggestions to correct grammar. It doesn’t come up with legal analysis on its own, and any suggestions for grammar changes should be scrutinized by the licensed professional to make sure the grammar changes don’t affect the argument.
And your second statement isn’t what happened here. If the lawyer had written an argument and then fed it to AI to correct and improve, then that would have the basis of starting with legal analysis written from a licensed professional. In this case, the lawyer bragged that he spent only seconds on this case instead of hours because the AI did everything. If he only spent seconds, then he very likely didn’t start the process with writing his own analysis and then feeding it to AI; and he likely didn’t review the analysis that was spit out by the AI.
This is an issue that is happening in the medical world, too. Young doctors and med students are feeding symptoms into AI and asking for a diagnosis. That is a legitimate thing to use AI for as long as the diagnosis that gets spit out is heavily scrutinized by a trained doctor. If they just immediately take the outputs from AI and apply the standard medical treatment for that without double checking whether the diagnosis makes sense, then that isn’t any better than me typing my symptoms into Google and looking at the results to diagnose myself.
Removed by mod
Now I’m imagining an AI lawyer in court, thanks.
That’s what the appeals process is for.
I don’t know about this particular lawyer, but I have heard that some lawyers will try novel court strategies, knowing that it’s a win-win situation. If the strategy works, then their clients benefit, and if the strategy doesn’t work, their clients get an appeal for having ineffective counsel where they normally wouldn’t have an appeal.
If a client gets an appeal for ineffective counsel how is that counsel not brought up before the bar for review? That seems like a death knell for a lawyer.
People don’t generally get sanctioned for making honest mistakes. I didn’t say that a lawyer would tank the case on purpose, just that they’d try a new strategy. If no lawyers were allowed to try new strategies without facing penalties, that also seems like a bad system.
Not sure if there is a procedure for when a lawyer is practicing, I have never heard of a bar referral after a ruling on a motion for ineffective assistance in NY, but I have heard of retiring attorneys landing on the grenade so to speak and writing affidavits claiming that anything they may have touched in the slightest was somehow deficient, spoiled or tainted by their involvement if it can get a shot at more billable work/appeal.
I had never heard that. Is there a name for that? Or do you have a place I can read more about it?
Well, the lawyer gave interviews after his client was guilty. Bragging about how instead of spending hours on it he only spent “seconds” and that the AI would mean he could have a lot more clients and make a lot more money.
So, it’s going to be pretty hard for him to now argue he put in just as much effort.
Did he help develop or train the AI? That upfront effort should perhaps be considered.
Removed by mod
Because he didn’t review it…
He used it “as is” so he could advertise his AI tool as “does it all by itself”.
It sounds like rather than advertising it as tool for lawyers, he’s advertising it to clients as a replacement for lawyers.
Removed by mod
In my opinion, how good the AI performed is irrelevant. What is is the fact that an AI was used instead of the lawyer.
If it is proven that the lawyer used what the AI delivered verbatim then it doesn’t matter how good that text was. The client has the right to have a lawyer, not an AI pretending to be a lawyer.
Removed by mod
Autocorrect does single words and you usually review each word. Something like ChatGPT will generate an entire document for you, it’s up to you if you want to verify the correctness of everything in there, which most people don’t.
Removed by mod
deleted by creator
Good point. If a lawyer is stupid enough to use AI, he’s probably too stupid to be a good lawyer in the first place
Removed by mod
In this case, it seems to be much worse, inventing reference cases and making nonsensical arguments.
https://youtu.be/oqSYljRYDEM
Here is an alternative Piped link(s):
https://piped.video/oqSYljRYDEM
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
That was mentioned briefly in the article. I was about to look into it a little more but I got side tracked. Thanks!