Here’s What Happens When Your Lawyer Uses ChatGPT

[Wargames]

If you make a custom GPT and upload the proper documents it will give legal word vomit you can edit into a decent statement of facts . You maybe can use it to evaluate other cases but you would still have to do a lot of writing to make sure the right cases are being quoted and everything forms the argument you want to make. Same for writing an application.

The issue is Lexis and Westlaw got all the data, and they arent dumb, they wont let the ai replace lawyers because lawyers pay them exorbitant fees. It takes just as much billable hours now as then, and is essentially a template builder.

It's not going to replace a lawyer anytime soon. I mean shyt I wouldn't fully trust it to create a case grid, let alone something I am putting in front of a judge. Like I said before you end up spending time editing instead of writing and like most AI I think that is the end result, now instead of writing briefs you have to edit them.

 

[bnew]

If you make a custom GPT and upload the proper documents it will give legal word vomit you can edit into a decent statement of facts . You maybe can use it to evaluate other cases but you would still have to do a lot of writing to make sure the right cases are being quoted and everything forms the argument you want to make. Same for writing an application.

The issue is Lexis and Westlaw got all the data, and they arent dumb, they wont let the ai replace lawyers because lawyers pay them exorbitant fees. It takes just as much billable hours now as then, and is essentially a template builder.

It's not going to replace a lawyer anytime soon. I mean shyt I wouldn't fully trust it to create a case grid, let alone something I am putting in front of a judge. Like I said before you end up spending time editing instead of writing and like most AI I think that is the end something instead of writing briefs you have to edit them.

are you using one type of model?

 

[Wargames]

are you using one type of model?

Unfortunately, I don't have the time to explore different models. It's not a word issue it's a content issue. Sure It can synthesize cases but can it apply multiple cases to various facts and then make the argument that overall this is why precedent says the rule should be interpreted in my clients favor? My whole brief is me just building a cohesive argument, using little passive aggressive things to try and reiterate my point of view and undercut what I think opposing counsel is trying to do. That is a lot to ask of a LLM.

You could break it up, use different prompts, review different results, and get what you want, but then you still got to use your own judgement and sew it all back together. It’s gonna take the same amount of time, and Westlaw and Lexis still gonna get their cut. Dont even get me started on the legal terms that LLM and spell check will autocorrect, but have actual legal meaning.

 

[bnew]

Unfortunately, I don't have the time to explore different models. It's not a word issue it's a content issue. Sure It can synthesize cases but can it apply multiple cases to various facts and then make the argument that overall this is why precedent says the rule should be interpreted in my clients favor? My whole brief is me just building a cohesive argument, using little passive aggressive things to try and reiterate my point of view and undercut what I think opposing counsel is trying to do. That is a lot to ask of a LLM.

You could break it up, use different prompts, review different results, and get what you want, but then you still got to use your own judgement and sew it all back together. It’s gonna take the same amount of time, and Westlaw and Lexis still gonna get their cut. Dont even get me started on the legal terms that LLM and spell check will autocorrect, but have actual legal meaning.

thats insightful

make use of double quotes for uncommon terms and even names. do you have some dummy data as an example or prompt you can share?

try a system prompt like "do not autocorrect or modify any words" or "do not autocorrect or modify any words, assume uncommon words are legal terms."

edit:
generated by clause 3 haiku
Prompt:
I am providing you with a legal document that contains specialized terminology, including Latin phrases. Please analyze the document, but do not attempt to correct, modify, or autocorrect any of the legal or Latin terms, even if they appear misspelled or unconventional. Preserve the original language and formatting of the document exactly as it is presented. Treat all Latin and legal phrases as intentional and do not make any changes to them. Provide your analysis while strictly maintaining the integrity of the legal terminology and formatting.

 

[Wargames]

thats insightful

make use of double quotes for uncommon terms and even names. do you have some dummy data as an example or prompt you can share?

try a system prompt like "do not autocorrect or modify any words" or "do not autocorrect or modify any words, assume uncommon words are legal terms."

Can you explain the double quotes for uncommon terms? I’m willing to give it a try. I still think the content/logic/subtle subversive aspects of legal writing is outside the scope but this could help.

 

[bnew]

Can you explain the double quotes for uncommon terms? I’m willing to give it a try. I still think the content/logic/subtle subversive aspects of legal writing is outside the scope but this could help.

use double quotes or even single quotes to be explicit about what you're referring to so it doesn't attempt to autocorrect anything and feed it additional context.

https://labs.perplexity.ai/

used the claude haiku model


you have to provide context and specify key terms and names in your prompts to reduce hallucinations.

edit:
partial screenshot of text

you don't have to use double quotes for the same word repeate4dly like I did above but once or twice should be enough to get the point across. you might want to look into find and replace browser extensions for the batch processing words in quotations.


I amended the other prompt:
I am providing you with a legal document that contains specialized terminology, including Latin phrases. Please analyze the document, but do not attempt to correct, modify, or autocorrect any of the legal or Latin terms, even if they appear misspelled or unconventional. Preserve the original language and formatting of the document exactly as it is presented. Treat all Latin and legal phrases as intentional and do not make any changes to them. Provide your analysis while strictly maintaining the integrity of the legal terminology and formatting. REPLY WITH: [To get started, please go ahead and provide the specific legal document(s) you would like me to analyze. I will be sure to follow the instructions you've outlined above and preserve the original language, even if it includes unusual legal or Latin terms.]

---

you would start off with that prompt and then after it replies, start chatting.

 

[bnew]

@Wargames see if you find this useful
created by claude3 haiku

Here is a more comprehensive, generic prompt for using an AI system like ChatGPT to assist with legal document drafting:

I am an attorney working on a [type of legal matter, e.g. civil litigation, corporate transaction, etc.] and I need your help drafting a [specific legal document, e.g. motion to dismiss, merger agreement, etc.].

Please review the following case files, legal documents, and examples that I have provided:
[List of relevant materials]

Using this information, I would like you to draft a [document type] that meets the following requirements:

Structure and Format:
- The document should be structured in the standard format and style expected by the [court/jurisdiction/industry standards], including appropriate sections, headings, and formatting.
- The tone and language should be professional, persuasive, and tailored to the intended audience (e.g. judge, opposing counsel, client, etc.).

Content Requirements:
1. Statement of Facts:
- Provide a detailed summary of the key facts and background information relevant to this matter, based on the provided case files.
- Ensure the facts are presented in a clear, logical, and compelling manner.

2. Legal Analysis:
- Identify the primary legal standards, precedents, and authorities that apply to this case or transaction.
- Summarize the key elements of the relevant law and how they relate to the facts of this matter.
- Outline the main legal arguments, reasoning, and conclusions that support the desired outcome.

3. Requested Relief/Outcome:
- Clearly state the specific legal action or outcome being sought (e.g. dismissal of claims, approval of merger, etc.).
- Explain how the presented facts and legal analysis justify the requested relief.

4. Additional Considerations:
- Flag any areas where you are unsure or believe human review and refinement would be beneficial.
- Explain your reasoning and decision-making process for the key elements of the document, such as the legal arguments, case law citations, and overall strategy.

Please provide an initial draft of the [document type] based on the instructions above. I will review your work and provide feedback. We can then collaborate to refine and finalize the document to ensure it meets all requirements and presents the strongest possible case.

Let me know if you have any other questions as you work on this project. I'm here to provide guidance and support throughout the drafting process.

OR​

Got it, here's an updated comprehensive prompt that requires the user to provide the relevant legal documents after the initial prompt:

"I am an attorney working on a [type of legal matter, e.g. civil litigation, corporate transaction, etc.] and I need your help drafting a [specific legal document, e.g. motion to dismiss, merger agreement, etc.].

Please provide the following case files, legal documents, and examples that I will need for this project:
[Prompt user to enter list of relevant materials]

Once you have provided those materials, I would like you to draft a [document type] that meets the following requirements:

Structure and Format:
- The document should be structured in the standard format and style expected by the [court/jurisdiction/industry standards], including appropriate sections, headings, and formatting.
- The tone and language should be professional, persuasive, and tailored to the intended audience (e.g. judge, opposing counsel, client, etc.).

Content Requirements:
1. Statement of Facts:
- Provide a detailed summary of the key facts and background information relevant to this matter, based on the provided case files.
- Ensure the facts are presented in a clear, logical, and compelling manner.

2. Legal Analysis:
- Identify the primary legal standards, precedents, and authorities that apply to this case or transaction.
- Summarize the key elements of the relevant law and how they relate to the facts of this matter.
- Outline the main legal arguments, reasoning, and conclusions that support the desired outcome.

3. Requested Relief/Outcome:
- Clearly state the specific legal action or outcome being sought (e.g. dismissal of claims, approval of merger, etc.).
- Explain how the presented facts and legal analysis justify the requested relief.

4. Additional Considerations:
- Flag any areas where you are unsure or believe human review and refinement would be beneficial.
- Explain your reasoning and decision-making process for the key elements of the document, such as the legal arguments, case law citations, and overall strategy.

Please provide an initial draft of the [document type] based on the instructions above and the materials you have provided. I will review your work and provide feedback. We can then collaborate to refine and finalize the document to ensure it meets all requirements and presents the strongest possible case.

---

Here is the updated prompt that requires the user to provide the necessary legal documents and materials:

Using the provided case files, legal documents, and examples, please draft a [type of legal document, e.g. motion to dismiss] for the [case name/details].

First, please provide the following case files, legal documents, and examples that I will need for this project:
[Prompt user to enter list of relevant materials]

Once you have provided those materials, please proceed with drafting the [document type] as follows:

1. Provide a detailed statement of facts based on the provided case information.

2. Summarize the key legal standards, precedents, and authorities that are most relevant to this case.

3. Outline the main arguments and reasoning for why [requested legal action, e.g. dismissal] is warranted, citing the applicable law and precedents.

4. Structure the document in the appropriate format and tone for submission to the [court/jurisdiction].

5. Highlight any areas where you are unsure or where you believe human review and refinement would be beneficial.

6. Explain your reasoning and decision-making process for the key elements of the document, such as the legal arguments, case law citations, and overall strategy.

Please provide this in draft form initially. I will review your work, provide feedback, and then ask you to revise and refine the document based on my input. The goal is to collaborate and iterate to produce a high-quality, legally sound [document type] that meets the court's requirements.

 

[bnew]

@Wargames

has LLM's improved for the tasks you give it?

 

[bnew]

Spoiler: thread continued

Spoiler: full text & large images

1/22
@JamesSurowiecki
Just figured out where these fake tariff rates come from. They didn't actually calculate tariff rates + non-tariff barriers, as they say they did. Instead, for every country, they just took our trade deficit with that country and divided it by the country's exports to us.

So we have a $17.9 billion trade deficit with Indonesia. Its exports to us are $28 billion. $17.9/$28 = 64%, which Trump claims is the tariff rate Indonesia charges us. What extraordinary nonsense this is.

[Quoted tweet]
It's also important to understand that the tariff rates that foreign countries are supposedly charging us are just made-up numbers. South Korea, with which we have a trade agreement, is not charging a 50% tariff on U.S. exports. Nor is the EU charging a 39% tariff.

2/22
@JamesSurowiecki
Even given that it's Trump, I cannot believe they said "We'll just divide the trade deficit by imports and tell people that's the tariff rate." And then they decided to set our tariffs by just cutting that totally made-up rate in half! This is so dumb and deceptive.



3/22
@JamesSurowiecki
How in God's name did Scott Bessent agree to sign off on this?



4/22
@JamesSurowiecki
This tweet is correct, but it's actually worse than I thought: in calculating the tariff rate, Trump's people only used the trade deficit in goods. So even though we run a trade surplus in services with the world, those exports don't count as far as Trump is concerned.
https://nitter.poast.org/JamesSurowiecki/status/1907559189234196942



5/22
@Psychonaut1001
I’m Australian,
America has a trade SURPLUS with us, and they’ve STILL put a 10% tariff on us.



6/22
@JamesSurowiecki
Yes, any country where we have a trade surplus or where the trade deficit/imports is less than 10% had their tariff rate listed as 10%.



7/22
@orthonormalist


[Quoted tweet]
DID I CRACK IT?

I think I figured out at least a chunk of the math.

It's trade deficit divided by their exports.

EU: exports 531.6, imports 333.4, deficit 198.2. 198.2/531.6 is 37, close to 39.

Israel: exports 22.2, imports 14.8, deficit 7.4. 7.4/22.2 is 33.


8/22
@JamesSurowiecki
They're also only looking at goods, so even though we run a massive services surplus with the world, that didn't count when they calculated these imaginary "tariff rates."



9/22
@mdangremond
How many calculations have you checked?



10/22
@JamesSurowiecki
A few. But it's obviously right. Look at the UK. The UK has a 20% VAT, which Trump says counts as a non-tariff barrier. But its tariff rate is supposedly only 10%. That's because our balance of trade with the UK is almost even, and all countries are listed as having at least a 10% tariff rate.



11/22
@SpencerHakimian
You cannot be serious.



12/22
@JamesSurowiecki
Amazingly, I am. They're also only looking at goods. So even though we run a massive services surplus with the world, Trump just ignored that when calculating the "tariff rate."



13/22
@JamesGozzi
Nonsense that is going to work.

Argentina - Removed all tariffs
Vietnam - Removed all tariffs
Israel - Removed all tariffs
Canada - Moving toward no tariffs
Mexico - No tariffs planned
India - Considering reducing tariffs
EU - Planning retaliation
Australia - No tariffs planned



14/22
@JamesSurowiecki
Do you anticipate American companies doing a lot of exporting to Vietnam now?



15/22
@Phili_Thomp
That's why some countries are charged less than their actual MFN rate

Turkey, 16.2%, 10% Trump rate
Argentina, 13.4%, 10% Trump rate
Morocco, 14%, 10% Trump rate

For what it's worth (nothing) there is a U.S.- FTA with Morocco... And it's the Mexico carmaker for Europe



16/22
@JamesSurowiecki
Yes - if our balance of trade with a country is roughly even (or trade deficit/imports is less than 10%), they're listed as 10%.



17/22
@ChrisKartalia
If the rates are inflated, all the easier for countries to “cut” them, Trump claims a win, and gets the 10% ongoing rate that he really wants



18/22
@JamesSurowiecki
In some cases, the rates listed are less than the rates countries actually charge, because we run a small trade deficit with them, and if trade deficit/imports is less than 10%, Trump just listed their tariff rate as 10%.



19/22
@0xKingAlex
This is pure cope. The chart includes tariffs and non-tariff barriers, VATs, currency manipulation, import bans, regs, etc.

You’re pretending it’s just deficit math to gaslight. It’s not. It’s reciprocity.

If they tax us 40%, we tax them 40%. Simple.



20/22
@JamesSurowiecki
Will you stop? They're not taxing us 40%. That's an imaginary number. The "tariff rates" in that chart are literally just our trade deficit in goods with a country/our imports from that country. They don't reflect actual tariffs + non-tariff barriers.



21/22
@wightmanr
Hrm... not disagreeing that the rates are fake / nonsense... but feels more arbitrary than suggested. EU for instance, the trade imbalance is only ~3% of the total trade if you factor in goods and services. The 20% seems closer to the VAT, which is also completely moronic to consider a tariff given that it applies to foreign and domestic goods equally... I think the average tariff rate in both directions up until now was about 1% in both directions.... regardless, WTF, where are the adults in the room?



22/22
@JamesSurowiecki
As I said later in the thread, they ignored services when calculating the "tariff rates." They only looked at goods.




To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196

Spoiler: full text & large images

1/11
@quendergeer
oh my fukking god

[Quoted tweet]
they've just asked chatgpt to spit out a spreadsheet haven't they

2/11
@JapanDeepValue1

3/11
@0xrooter
they forgot to turn on deep research



4/11
@thisiswenzel
Haha, I'm dead.



5/11
@georgia_genx
And now Musk is leaving... lol



6/11
@Rafi_0x
We’re 100% living in a simulation, there’s no other explanation



7/11
@SantiagoPombo
I thought I was kidding:

[Quoted tweet]
Liberation Day Tariffs, Tech Edition:
Vibe coders seized the global codebase, rewrote the economy in one untested commit with zero coverage and no rollback plan, via "git push --force" to main. Servers are melting.

README: "Trust the vibes"


8/11
@stoitz
So funny, I was wondering why they had separate tariffs on St Pierre and Miquelon



9/11
@ISpotLies
The island populated by penguiks is owned by Australia and to avoid tariffs Australia could just send exports through the island. That's why it was listed actually.



10/11
@RyanPMcGowan
Low info post masquerading as high info.



11/11
@xDreamCoding
Vibe governing




To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196

Spoiler: text

Commented on Thu Apr 3 19:13:33 2025 UTC

You can tell they used Ai , they put tarrifs on islands that don’t even have people on them.

│
│

Spoiler: text

│ Commented on Thu Apr 3 20:05:32 2025 UTC
│
│ Stable Genius Trump Just Put Tariffs on a U.S. Military Base.
│

 

[Rev Leon Lonnie Love]

Buu buu but AI gonna take all our jobs

 

[bnew]

Buu buu but AI gonna take all our jobs

it will but it's a tool right now and if you don't know how to use the tool to produce results you want it's not entirely the fault of said tool.

garbage in, garbage out.

 

[Rev Leon Lonnie Love]

it will but it's a tool right now and if you don't know how to use the tool to produce results you want it's not entirely the fault of said tool.

garbage in, garbage out.

that shyt will never happen

 

[bnew]

Judge slams lawyers for ‘bogus AI-generated research’

The law firms involved must pay $31,000.

www.theverge.com

Judge slams lawyers for ‘bogus AI-generated research’​

Judge Michael Wilner said ‘no reasonably competent attorney’ should outsource research to AI.
by Emma Roth

May 13, 2025, 6:22 PM EDT
7 Comments7 New

Illustration by Cath Virginia / The Verge | Photos from Getty Images
Emma Roth is a news writer who covers the streaming wars, consumer tech, crypto, social media, and much more. Previously, she was a writer and editor at MUO.

A California judge slammed a pair of law firms for the undisclosed use of AI after he received a supplemental brief with “numerous false, inaccurate, and misleading legal citations and quotations.” In a ruling submitted last week, Judge Michael Wilner imposed $31,000 in sanctions against the law firms involved, saying “no reasonably competent attorney should out-source research and writing” to AI, as pointed out by law professors Eric Goldman and Blake Reid on Bluesky.
“I read their brief, was persuaded (or at least intrigued) by the authorities that they cited, and looked up the decisions to learn more about them – only to find that they didn’t exist,” Judge Milner writes. “That’s scary. It almost led to the scarier outcome (from my perspective) of including those bogus materials in a judicial order.”

As noted in the filing, a plaintiff’s legal representative for a civil lawsuit against State Farm used AI to generate an outline for a supplemental brief. However, this outline contained “bogus AI-generated research” when it was sent to a separate law firm, K&L Gates, which added the information to a brief. “No attorney or staff member at either firm apparently cite-checked or otherwise reviewed that research before filing the brief,” Judge Milner writes.

When Judge Milner reviewed the brief, he found that “at least two of the authorities cited do not exist at all.” After asking K&L Gates for clarification, the firm resubmitted the brief, which Judge Milner said contained “considerably more made-up citations and quotations beyond the two initial errors.” He then issued an Order to Show Cause, resulting in lawyers giving sworn statements that confirm the use of AI. The lawyer who created the outline admitted to using Google Gemini, as well as the AI legal research tools in Westlaw Precision with CoCounsel.

Related​

• Judge berates AI entrepreneur for using a generated ‘lawyer’ in court

This isn’t the first time lawyers have been caught using AI in the courtroom. Former Trump lawyer Michael Cohen cited made-up court cases in a legal document after mistaking Google Gemini, then called Bard, as “a super-charged search engine” rather than an AI chatbot. A judge also found that lawyers suing a Colombian airline included a slew of phony cases generated by ChatGPT in their brief.
“The initial, undisclosed use of AI products to generate the first draft of the brief was flat-out wrong,” Judge Milner writes. “And sending that material to other lawyers without disclosing its sketchy AI origins realistically put those professionals in harm’s way.”

 

[bnew]

Eric Goldman (@ericgoldman.bsky.social)

Angry judge roasts Biglaw lawyers for their "collective debacle" filing a brief where "~9 of the 27 legal citations in the 10-page brief were incorrect in some way" due to Generative AI https://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=3920&context=historical Lawyers will pay $31k...

bsky.app

Spoiler: full text & large images

1/36
ericgoldman.bsky.social
Angry judge roasts Biglaw lawyers for their "collective debacle" filing a brief where "~9 of the 27 legal citations in the 10-page brief were incorrect in some way" due to Generative AI

digitalcommons.law.scu.edu/cgi/viewcont...

Lawyers will pay $31k for their sloppiness
https://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=3920&context=historical

2/36
kasper11.bsky.social
I really wish the special master demanded that the plaintiff's attorneys produce their billing records related to the motion. Really, really curious how much time they billed for "research" and preparing the outline.

3/36
bob021341.bsky.social
Exactly. We know the bill would have been huge had this not been caught. The affidavit of the associate at K&L describes his role in writing the brief and admits that he did not check the cites provided by the first firm; he apparently only did his own supplemental research.

4/36
aftd.dev
It's bizarre, in this day and age, that lawyers have automated drafting a brief but not checking the citations.

I've been out of practice for a while, but Lexis and WestLaw make automated, non-LLM citation checkers right? Do they not work?

5/36
cfit.bsky.social
I was selling Westlaw's automatic citation checker over a decade ago, and the big law firms all had it. No excuse for this.

6/36
nywonk.bsky.social
They should pay with their law license.

That would be a sufficient deterrent. After the first ambulance chaser gets disbarred, the rest will think twice before using AI.

7/36
peanut24.bsky.social
Great decision here but I wish the Special Master hadn't been so reticent about "naming and shaming" the lawyers involved. It seems they had several opportunities to correct the problems, and didn't. Really sloppy lawyering, all around.

8/36
bob021341.bsky.social
They're not named in the Order, but their identities are revealed in their affidavits appended to the Order wherein they tried to explain away their conduct.

9/36
christinepnw.bsky.social
It's curious that the judge is actually the one who figured this out, and not opposing counsel in reviewing the opposing brief. As an attorney, it's satisfying to be able to call out the opposition in a responsive brief when they misstate arguments, much less flat out inventing citations.

10/36
lu.is
The law *firms* will pay for the lawyer’s sloppiness. The judge concludes no individual lawyers should be sanctioned, because as ever we protect our own interests over those of our clients.

11/36
adistantdream.bsky.social
I dunno, $31K might as well be ¢31 to a biglaw firm. $31M might make them look twice.

12/36
dewtell.bsky.social
There were in-case sanctions as well (paragraph 20): "I have struck, and decline to consider, any of the supplemental briefs ... If the undisclosed use of AI ... causes a client to lose a motion or case, lawyers will undoubtedly be deterred from going down that pointless route."

13/36
dewtell.bsky.social
I can't imagine the firm's insurance carrier is too happy about that language. The idea that the next time this shyt happens it could result in a slam-dunk multi-million $ malpractice claim instead of just losing a privilege argument this time is probably enough to get the firm's attention. 2/n

14/36
writesomethink.bsky.social
Because at its core, AI is garbage.

15/36
ubermitch.bsky.social
lol, and the special master they tried to pull this in front of is former Magistrate Judge Wilner; as an occasional practitioner before CD Cal, he's *precisely* someone you don't want to piss off

16/36
dirtside.bsky.social
but biglaw firms are respectable professionals, biglaw firms have told me! could it be that being big and powerful doesn't equate to being actually good at your job?!

17/36
roncampbell.bsky.social
This is a very strongly worded order. Here's a sample from one footnote: "I consider those errors to be at the mild end of the AI hallucination spectrum." The special master did not come to play.

18/36
dabenner.bsky.social
@clapifyoulikeme.favrd.social

19/36
clapifyoulikeme.favrd.social
Oh Boy

20/36
petitecarcajou.bsky.social
And this is why you always follow up with non-AI products to check your research before uploading documents to the court. (Along with spelling / grammar / formatting check and ensuring that you’re following any special court or chambers’ rules.)

21/36
vanfirth.bsky.social
They got off relatively lightly, no individual sanctions and a smallish penalty. Presumably because they confessed their sins and it wasn’t malfeasance, just utter blithering stupidity. “Go and sin no more.”

22/36
wut12345.bsky.social
But here’s a small point: Ellis George isn’t considered BigLaw. They don’t even have 100 attorneys. They’re a mid sized firm and that’s exactly where I would expect to see these shenanigans

23/36
michiforjoy.bsky.social
Ellis George sent a draft portion of the brief to K&L Gates co-counsel, who it seems failed to check anything.

24/36
timack.bsky.social
A bit off-topic here, but the judge has very clear and easily-read writing style.

25/36
sybilltrelawney.bsky.social
Why do lawyers keep doing this?

26/36
riverofdiversity.bsky.social
So they can charge for their time, while doing other things.

27/36
viewfromarchstreet.com
I vote for $15K per incorrect AI citation after first offense

28/36
rantlaw.bsky.social
The frustrating thing for me is that the source of the errors is unclear. Was it Gemini or the commercial LLM? I suspect at least some came in through the commercial tool but crave evidence.

29/36
wachob.com
Imagine if instead of using un-ready generative AI, they were investing in making AI a better *tool*.

I am in the "AI is coming, let's deal with it" camp, so my attitude is "make the tool better" instead of just rolling my eyes at people using it.

That being said, <eyeroll> for not checking work

30/36
umlauthuth.bsky.social
They need to be debarred.

31/36
einsofasectional.bsky.social
they should be disbarred.

32/36
faeelin.bsky.social
What firm

33/36
brianstorms.bsky.social
Clearly it is time for AI-based judges.

34/36
ihorgowda.bsky.social
@garymarcus.bsky.social

35/36
mad-pat.bsky.social


36/36
phillyphile215.bsky.social


To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196