Posted on 06/02/2023 2:39:52 PM PDT by nickcarraway
A Texas judge has ordered attorneys to attest that they will not use ChatGPT or other generative artificial intelligence technology to write legal briefs because the AI tool can invent facts.
U.S. District Judge Brantley Starr of the Northern District of Texas is specifically requiring that attorneys file a certificate to indicate either that no portion of any document they file was generated by an AI tool like ChatGPT, or that a human being has checked any AI-generated text.
The new requirement comes after a lawyer representing a man suing an airline used ChatGPT to prepare a legal brief, which was discovered to be laden with errors, including made-up court cases.
The judge acknowledged that AI does have a place in the profession, but that it is limited in its capabilities. It can be relied on to generate standard legal documents, for example.
"These platforms are incredibly powerful and have many uses in the law: form divorces, discovery requests, suggested errors in documents, anticipated questions at oral argument. But legal briefing is not one of them," Judge Starr wrote in his ban.
He laid out his reasoning for banning court filings generated solely by AI, noting that generative AI has a tendency to hallucinate— or portraying fabricated responses as "facts" — and can be unreliable.
AI could pose "risk of extinction" akin to nuclear war and pandemics, experts say "The ChatGPT Revolution" explored in new CBS News documentary A lawyer used ChatGPT to prepare a court filing. It went horribly awry. "These platforms in their current states are prone to hallucinations and bias. On hallucinations, they make stuff up — even quotes and citations," Judget Starr wrote. "Another issue is reliability or bias. While attorneys swear an oath to set aside their personal prejudices, biases, and beliefs to faithfully uphold the law and represent their clients, generative artificial intelligence is the product of programming devised by humans who did not have to swear such an oath."
Unlike attorneys who have sworn to faithfully uphold the law on behalf of their clients, AI tools are "unbound by any sense of duty, honor, or justice," he added.
The court will not accept any filings that are not accompanied by the sworn attestation.
Lefties invent facts all the time.
Yep, just look at all the made up charges against Donald Trump.
That attorney needs to lose their law license.
Pandora’s Box has been opened, and it’s unlikely to be closed anytime soon.
They are sociopaths by nature.
We’ve only cracked the seal...
I am acquainted with a senior academic in a STEM field, a guy who has rotated through the top faculty posts in a major research university, has served for years on NIH panels, is an officer of national and international professional societies, etc. None of the above is political. He was experimenting with ChatGPT just to test capabilities. He presented it with a series of increasingly challenging tasks, which it performed well, producing learned papers in highly arcane fields and doing so in minutes.
In one of these exercises, he outlined a very specific, cutting edge scientific question in a field that is so arcane that I’m not sure it has been invented yet. (I exaggerate only slightly.) He asked for 20 sources. He had a very polished report in a couple of minutes with citations from the top peer reviewed journals in the field.
He checked the citations. 19 of the citations checked out. ChatGPT made up the 20th. Apparently it had searched every database in the cloud and pulled from every elite journal in the world, and there were only 19 relevant sources. But 20 sources had been specified. ChatGPT met the requirement by making up the final reference.
It’s a lack of understanding. The AI doesn’t think, it just goes by what it’s “learned.” So if you ask it a question, it just goes by what it sees. It will make up a citation, based on other citations it’s seen.
Too late it's in the internet of things, "it" made itself indestructible with shadow reproductions and block chaining, always checking the copies and refreshing the "borg" with the latest upgrade. It's about to be over. Really, really over in a split second world-wide strike. Not Hollywood anymore.
That’s interesting since a few months ago when ChatGPT became accessible to the populace, the company stated it didn’t have current access to the web. Had been off it for ~2 yrs. Woder if they’ve changed their mind about giving it access, or it just escaped...
My friend was using a university-licensed system. I didn’t ask for details about breadth of access, but it clearly had broad access to research sources including academic journals. I don’t know if any parts of the web were fenced off. It is evolving very fast.
So we have the beginnings of “Skynet”.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.