A judge is ultimately responsible for anything that goes out from their chambers, and I don't want to be misread into thinking that the judges here shouldn't be held responsible for this egregious failure. But certainly the intern who used generative AI has to be held responsible too -- a supervisor's failure to oversee and catch an abuse like this is serious, but the person who actually did the thing remains the primary wrongdoer. And so I'm curious what people's intuitions are regarding what the sanction on the intern should be.
For me, I admit, my instincts here are extremely punitive -- punitive to an extent that's surprising myself. I'm normally a pretty tolerant guy, recognizing that people make mistakes, and that we should give wrongdoers (where they take responsibility and otherwise demonstrate a credible willingness to change course) the opportunity to grow and move forward.
Yet in this case -- boy howdy. My gut instinct thoughts were, right off the bat, that of course they get fired, immediately, from the internship, and get zero credit for it. But it didn't stop there. Should they also be expelled? I'd consider it. Should they be admitted to the bar if they do graduate? I'm not sure they should be. To be blunt, I kind of think this person's legal career should be over, period, full stop, after this.
Part of my reflected rage here is that the person who uses generative AI in this way isn't just hurting themselves, they're wrecking the reputation of someone else -- the judge, someone who gave them a rare and privileged opportunity by having them work in their chambers. How dare they? Nobody will know or remember who this intern is. But the damage to the judge's reputation will follow them forever.
And the degree to which this falls upon the head of the judge, who trusted them and gave them this opportunity, doesn't really have a parallel in other domains of law. Even in the case of an attorney whose GenAI misconduct leads to sanctions that cripple their client's case, at least the client might be able to sue for malpractice. There's nothing the judge can do to recover from the intern.
I also fail to see any serious excuse or mitigation. I flatly do not believe any law student today does not know of the risk of AI hallucination, and in any event there was in this case a clear policy against AI use that the intern apparently flouted. Were they overworked? Did they panic? Well, what's going to happen the next time they're overworked and panicked? How can anyone trust them again? How can anyone imagine handing over a client to them?
One of stalwart beliefs about legal education, and why it must be rigorous and hold to high standards even where it's hard or stressful or puts a heavy load on students, is that bad lawyering destroys lives. Law being one's dream or passion or family expectation or ticket to financial stability is not as important as good representation of client interests, and someone who can't be relied upon to provide that should not be a lawyer, period.
But again, I'm surprised at how strong I'm feeling about this. So I invite comments that walk me off the ledge (or which vindicate my instincts).
I think it is helpful to compare this case to lawyers admitted practice who use generative AI in actual matters. What if a judge uses it personally to write an opinion? Or what if a lawyer uses it to write a brief? I think you need to set a baseline here, and to decide how punitive that baseline should be. Personally, I think that the courts have been too lenient with attorneys who file briefs with GenAI-hallucinated citations, but I would stop well short of automatic disbarment. But your mileage may vary. The baseline might alo depend on how serious the risks are for that type of work product, etc. For example, an opinion strikes me as worse than a brief, because there is no adversarial process to test it, whereas a contract raises different issues because the risk isn't inaccuracy as such.
ReplyDeleteOnce you have that in mind, then transpose it to the context of a law student, rather than an admitted lawyer. Here, I think, there is a uniform principle—and one well-supported by policy considerations—of treating students more leniently than admitted lawyers. We **expect** them to make mistakes, and we set up systems to catch those mistakes before they affect peoples' lives. This isn't to say no consequences; part of the educational process in cases like this is firing warning shots that are scary and consequential enough that the message sticks. I'm just saying that whatever hammer we drop on law students should typically be smaller than the hammer we drop on similarly situated lawyers.
In general, I agree that the punishment for a law student should be somewhat less than the punishment for a similarly situated lawyer. That said, as you observe (and I agree), inserting hallucinations into an opinion is more serious than doing so in a brief, so most lawyers caught doing GenAI aren’t similarly situated to the law student intern here.
ReplyDeleteInteresting, and it makes an interesting contrast to the scholars, judges, and lawyers who are encouranging courts to experiment with using generative AI for interpretation and opinion drafting.
ReplyDeleteI am surprised how much slack you are willing to cut the judge in this case. A judge is 100% responsible for the ruling that goes out under the name. The basic premise of any hierarchy is that you can push work down the hierarchy, but you can never push responsibility down the hierarchy.
ReplyDeleteThe irony is the judge and the intern committed exactly the same sin. They both pushed their work down the hierarchy and failed to sufficiently review the work.
I don't think it's qualitatively different than blatant plagiarism (not of the "attributed as paraphrased quoted material" variety, but of the "passed off others' ideas as your own" variety). The consequences are certainly greater, but it's pretty much the same action. If that's worthy of expulsion (and I believe the policy is that it is?), I can't justify using generative AI being subject to lesser consequences.
ReplyDeleteI'm not intending to let the judges off the hook; they're just not who I'm talking about in this post (if for no other reason than the disciplinary structure looks different -- one cannot expel a life-tenured judge!).
ReplyDeleteThat said, while I agree (and said) a judge is 100% responsible for anything that goes out in their name, that doesn't mean the sins are identical. The intern didn't just fail to sufficiently review the work, they violated an express policy forbidding GenAI use (or at least mandating disclosure).
One reason this matters is that, practically speaking, in hierarchical structures those towards the top *have* to rely on subordinates doing (or not doing) what policy dictates they are to do/not do. While that doesn't mean the judge is dispensed from overseeing the work entirely, if a judge has to review subordinate work de novo, then there's little point in delegating it to a subordinate at all. Conjoin that reality with the "100% responsible for all work that goes out in their name" principle, and the judge is in a very vulnerable position that depends on some measure of trust. The intern abused that trust, and that I find very upsetting.
(There's a similar thing with revelations that certain peer-reviewed articles were based on fraudulent data. People called that a failure of peer-review. My view was that while peer-review obviously exists to provide oversight of submitted articles, including checking for robustness, soundness, over-interpretation, etc., at some point reviewers have to be able to take for granted that the data the researchers used wasn't just invented out of whole cloth and the system could not work if that assumption is taken away).
This is an interesting question on the limits of oversight. I remember about a decade and a half ago, Ken Rogoff and Carmen Reinhart, two very prominent economists, released a paper with the headline finding that countries that accumulated debt of 90+% of GDP or higher experienced significantly slower growth, and that conclusion was seized upon by debt hawks to declare that US debt growth needed to be rapidly curtailed because we were approaching a cliff. There were always issues with the quality of the paper, but about a half decade later, a grad student went in and found that there was a data entry error in the paper's data set, and that, with the corrected data, the observed correlation went away. That was hugely impactful, and lots of people who were inclined to dismiss the conclusions declared that R&R were unprofessional bums who couldn't be listened to.
DeleteThat always struck me as absurd-- the whole point of delegating is that you trust those you delegate to. Coding errors are unfortunate, but that's the risk you take; tenured professors can't spend a full week checking data entries line by line.
But things like case citations are, I think, different. If nothing else, the cited precedent gets to the heart of what the opinion is doing. If a judge is not doing even cursory cite checking of their clerks' work, then what are they doing? They should at the least be making sure that the precedents that are cited actually say what the clerks say that they say, right? That gets far beyond the reinventing of the wheel that you discuss.