"Artificial intelligence (AI)" and the law thread 🧵
-
"Artificial intelligence (AI)" and the law thread 🧵
(Really large language models (LLM), not AI, but I'll never find the thread if I don't put "AI" in it)
-
"Artificial intelligence (AI)" and the law thread 🧵
(Really large language models (LLM), not AI, but I'll never find the thread if I don't put "AI" in it)
-
BC lawyer sanctioned for hallucinated cases.
"The cases never made it into Ke's arguments; they were withdrawn once she learned they were non-existent."
https://www.cbc.ca/news/canada/british-columbia/lawyer-chatgpt-fake-precedent-1.7126393
-
BC lawyer sanctioned for hallucinated cases.
"The cases never made it into Ke's arguments; they were withdrawn once she learned they were non-existent."
https://www.cbc.ca/news/canada/british-columbia/lawyer-chatgpt-fake-precedent-1.7126393
There are some things where you say to yourself, "I'm not sure how I feel about this."
This is not one of those things.
An AI video of a victim was presented at a criminal sentencing hearing to make a victim impact statement.
https://www.nbcnews.com/news/us-news/road-rage-victim-speaks-killers-sentencing-rcna205454
-
There are some things where you say to yourself, "I'm not sure how I feel about this."
This is not one of those things.
An AI video of a victim was presented at a criminal sentencing hearing to make a victim impact statement.
https://www.nbcnews.com/news/us-news/road-rage-victim-speaks-killers-sentencing-rcna205454
Guy tried to play a video of an avatar to present his arguments. The judges immediately turned it off because they hadn't been told that this is was what was being submitted.
https://apnews.com/article/artificial-intelligence-ai-courts-nyc-5c97cba3f3757d9ab3c2e5840127f765
-
Guy tried to play a video of an avatar to present his arguments. The judges immediately turned it off because they hadn't been told that this is was what was being submitted.
https://apnews.com/article/artificial-intelligence-ai-courts-nyc-5c97cba3f3757d9ab3c2e5840127f765
Uh oh another attorney getting ordered by the court to appear and show cause why he should not be sanctioned for fake cases in his memorandum of law. (Because he used "AI" to draft it.)
Here's the court order:
https://storage.courtlistener.com/recap/gov.uscourts.nced.221403/gov.uscourts.nced.221403.119.0_1.pdf -
Uh oh another attorney getting ordered by the court to appear and show cause why he should not be sanctioned for fake cases in his memorandum of law. (Because he used "AI" to draft it.)
Here's the court order:
https://storage.courtlistener.com/recap/gov.uscourts.nced.221403/gov.uscourts.nced.221403.119.0_1.pdf"In his reply, Fivehouse asserted that the response brief contained fabricated quotations and misstated the holdings of several cases.
The court directed Renfer to respond.
...
-
"In his reply, Fivehouse asserted that the response brief contained fabricated quotations and misstated the holdings of several cases.
The court directed Renfer to respond.
...
... In a surreply, he stated that he “inadvertently included incorrect citations to case law from this Circuit,” attributing the errors to the “inadvertent filing of an unfinalized draft document.”
...
Having reviewed the filings in this matter and other submissions by Renfer, the court has serious concerns about the accuracy of certain quotations and representations in Renfer’s filings,
and the explanation offered for their inclusion."Good for the court for saying "and the explanation"
-
... In a surreply, he stated that he “inadvertently included incorrect citations to case law from this Circuit,” attributing the errors to the “inadvertent filing of an unfinalized draft document.”
...
Having reviewed the filings in this matter and other submissions by Renfer, the court has serious concerns about the accuracy of certain quotations and representations in Renfer’s filings,
and the explanation offered for their inclusion."Good for the court for saying "and the explanation"
Trying CanLII's stupid "AI" search/analysis.
My FIRST result was bad.
First, the "AI" result:
"Hyundai Auto Canada Corp. v Engen, 2023 ABCA 85. The decision gives the appeal outcome (allowed in part), but it does not state any motion result."
Wrong. The VERY FIRST LINE of that decision:
"[1] The appellants (collectively, Hyundai) appeal the certification judge’s decision certifying the product liability action of the respondent, Robert Engen, as a class proceeding"
-
Trying CanLII's stupid "AI" search/analysis.
My FIRST result was bad.
First, the "AI" result:
"Hyundai Auto Canada Corp. v Engen, 2023 ABCA 85. The decision gives the appeal outcome (allowed in part), but it does not state any motion result."
Wrong. The VERY FIRST LINE of that decision:
"[1] The appellants (collectively, Hyundai) appeal the certification judge’s decision certifying the product liability action of the respondent, Robert Engen, as a class proceeding"
AI summary of oppression remedy in England says creditors have standing to bring such an action. It cites two sources.
Neither source says that.
🤷
-
AI summary of oppression remedy in England says creditors have standing to bring such an action. It cites two sources.
Neither source says that.
🤷
Another day another client hands me AI slop to review. Sigh.
-
_CEO Asks ChatGPT How to Void $250 Million Contract, Ignores His Lawyers, Loses Terribly in Court_
https://www.404media.co/ceo-ignores-lawyers-asks-chatgpt-how-to-void-250-million-contract-loses-terribly-in-court/ -
_CEO Asks ChatGPT How to Void $250 Million Contract, Ignores His Lawyers, Loses Terribly in Court_
https://www.404media.co/ceo-ignores-lawyers-asks-chatgpt-how-to-void-250-million-contract-loses-terribly-in-court/Third case about a lawyer reprimanded for fake cases in their factum, likely because of the use of artificial intelligence.
https://www.crookston.ca/wp/2026/03/12/false-quotes-fake-citations-and-artificial-intelligence/
-
Third case about a lawyer reprimanded for fake cases in their factum, likely because of the use of artificial intelligence.
https://www.crookston.ca/wp/2026/03/12/false-quotes-fake-citations-and-artificial-intelligence/
Added a fourth to that page.
Zhang v Chen, 2024 BCSC 285.
-
Wait a second, I may have just taken on a second job.
i just searched "artificial intelligence" and hallucination, and 138 results popped up.
The majority are not articles about the topic but actual cases where the decision maker has made comments that the cases before it are hallucinated.
Edit: Upon review, the vast majority are self-represented people using AI. That they do so is not helpful, but at least understandable. I won't be focussing on these cases here.
-
Wait a second, I may have just taken on a second job.
i just searched "artificial intelligence" and hallucination, and 138 results popped up.
The majority are not articles about the topic but actual cases where the decision maker has made comments that the cases before it are hallucinated.
Edit: Upon review, the vast majority are self-represented people using AI. That they do so is not helpful, but at least understandable. I won't be focussing on these cases here.
It is very clear from the sheer volume of such decisions that the courts and tribunals are getting absolutely PUMMELLED with this garbage.
-
It is very clear from the sheer volume of such decisions that the courts and tribunals are getting absolutely PUMMELLED with this garbage.
A federal judge just disqualified both parties' lawyers from practicing before the courts for 2 years. They were also each fined about $3,000.
Each side had used "artificial intelligence" to generate their materials and the materials were full of fake cases. Good times.
-
A federal judge just disqualified both parties' lawyers from practicing before the courts for 2 years. They were also each fined about $3,000.
Each side had used "artificial intelligence" to generate their materials and the materials were full of fake cases. Good times.
"Aycock barred the lawyers on each side who either used AI for legal research or for drafting the filings at issue — Kathleen Wilson and Kathryn Williams — from appearing in the Northern District of Mississippi for two years. Wilson was fined $2,500, and Williams was fined $3,500."
-
"Aycock barred the lawyers on each side who either used AI for legal research or for drafting the filings at issue — Kathleen Wilson and Kathryn Williams — from appearing in the Northern District of Mississippi for two years. Wilson was fined $2,500, and Williams was fined $3,500."
We just tested* a legal LLM and at least about 1 3rd of the cases it returned were not relevant to the question or the cases did not exist.
I tested it filing in citations only and it seemed to do that reasonably well but it still got two wrong. This was just looking up the actual citations of a case with a given name.
(Edit: *I should clarify that this thing came bundled in another service we already have. So this is not a service we sought out. It was shoved into my face.)
-
We just tested* a legal LLM and at least about 1 3rd of the cases it returned were not relevant to the question or the cases did not exist.
I tested it filing in citations only and it seemed to do that reasonably well but it still got two wrong. This was just looking up the actual citations of a case with a given name.
(Edit: *I should clarify that this thing came bundled in another service we already have. So this is not a service we sought out. It was shoved into my face.)
This appears to be a huge waste of time and money and energy and this isn't anything I would pay for.
I think I've pointed out elsewhere that when I ran a search in the AI search box versus the regular search box, I basically got the same results back.
I presume the tests that they run these things against are tailored to produce great results and that's how they conclude that they're operating at a PhD level. Because it's giving me garbage.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login