I attended the hearing for the hapless lawyer who used ChatGPT in his filings and here's my report. A cautionary tale not just for lawyers but for journalists: not about the machine but about our responsibility.
@jeffjarvis ´ChatGPT, as it is a wronged party in this case: wronged by the lawyers and their blame, wronged by the media and their misrepresentations, wronged by the companies —Microsoft especially — that are trying to tell users just what Schwartz wrongly assumed: that ChatGPT is a search engine that can supply facts. It can’t. It supplies credible-sounding — but not credible — language. That is what it is designed to do. That is what it does, quite amazingly. ´
@jeffjarvis
"Judge Castel’s point stands: It was the lawyer’s responsibility — to themselves, their client, the court, and truth itself — to check the machine’s work. This is not a tale of technology’s failures but of humans’, as most are."
@jeffjarvis you're right that much more ridicule will be deserved by the next hapless lawyer falling for the same trap than the first one. That is not a sufficient excuse though - a difference between a professional and someone who is merely paid to do some semblance of a job is that a professional takes responsibility for their work. This Schwartz character obviously did not.
@osma@jeffjarvis It’s all bs. I cannot imagine citing a case I did not first read and “Shepardize” to make sure it was good law. And, especially, if I were submitting it under a colleague’s name in a federal jurisdiction to which I was not admitted to practice.
@osma@jeffjarvis I think you overestimate the amount of actual research work that a lawyer with 30 years of experience normally does themselves, as opposed to relying on paralegal staff.
@jeffjarvis Well written! I'm glad to see some emphasis being put on the lawyers' responsibility to fact-check their research and their negligence in not doing anything other than going through ChatGPT to verify it. A lot of the coverage I've seen (not that it's very much) seems to want to blame the technology and ignore the blatant misuse of it by humans who should know better, or at least should have learned better at some point in the process.
@jeffjarvis THIS - good journalism is not the ability to write a story though writing it well will help with getting the story consumed in the end - good journalism is the ability to find and research and explain what is going on, in the world, in a situation, with people.
@jeffjarvis The part that bothers me the most is... what exactly did these lawyers think their clients were paying them for? If they really thought that just letting ChatGPT do their job for them was legitimate, why wouldn't their clients just do that themselves instead of paying a lawyer an absurd amount of money to do it for them? That alone seems like it should be enough to prove dereliction of duty.
@LouisIngenthron@jeffjarvis
This exactely.
In the age of the internet pretty much everything can be looked up online. What people pay experts for is to do that search on their behalf and filter out the relevant information amidst all the piles of bullshit and almost-but-not-quite-the-same
@jeffjarvis It's wild to see everyone realize that you must verify your sources/citations, and in some cases (e.g. journalism) get multiple sources when the evidence is weak (e.g. hearsay vs video proof*), or otherwise validate the data yourself, e.g. when someone claims "X is vulnerable to a security vuln" we would simply ask for the reproducer/affected lines of code and then validate it ourself, wild, I know.
A lot of people and industries have been coasting for a long time on "it's mostly probably correct so why bother checking? I have enough work as it already is..."
What's wild to me is that ChatGPT is (now, with the web plugins and especially PDF plugins) really good at checking sources, e.g. "go get this URL/document and create a summary, does it at all support X like the author is claiming", e.g. I had some citations that were to legitimately existing https://arxiv.org/ papers, but had nothing to do with what was being discussed.
(* Although now in the era of deep fakes... yeah. I got no idea. Good luck).
Add comment