AI hallucinations now include fake witnesses and testimony

 Above the Law is reporting that New Mexico attorney Stephen Aarons was held in contempt and fined $5,000 after filing an appellate brief containing “false testimony from wholly fabricated witnesses."  This is a different type of "hallucination" which -- at least to my knowledge - had been limited to made up citations or quotes.  You can read the story here.

Article on Lawyers’ Use of AI Notetakers and Client Conversations: Ethics Rules, Risks, and Best Practices

 Would using an AI assistant to take notes of a conversation with a client compromise attorney-client privilege or the duty of confidentiality?   Carolyn Elefant has a short article addressing the issue here and a similar "ethics opinion" here.

Two judicial circuits in Florida adopt rule on use of AI

 Two of Florida’s largest judicial circuits now require lawyers and self-represented litigants to disclose AI-generated court submissions, and to certify their accuracy, but according to at least one commentator the new rule essentially just "reminds lawyers of something that predates computers, the internet, and probably most of the courthouses we practice in today: if you put your name on a filing, you are responsible for what is in it."  For more on the story go here and here.

It is true that the new rule states the obvious: that it is required that all AI-generated content must be reviewed for accuracy.  But the new rule goes further by requiring actual disclosure of the use of AI, something I don't think many other jurisdictions are requiring explicitly yet.

According to one source, the order states that “[a]ny attorney or self-represented litigant who uses any generative artificial intelligence tool in the preparation of a pleading, motion, memorandum, response, proposed order, or other court record, must disclose such use on the face of the filing.”

Interestingly, the two circuits apparently did not agree on the appropriate sanctions for violations of the rule.

Failure to comply with the order in the 11th Judicial Circuit could lead to sanctions that include, “striking of the filing; denial of requested relief; monetary sanctions; contempt proceedings; referral to The Florida Bar or other appropriate authority; and any other sanction deemed appropriate by the Court.”

Failure to comply with the order in the 17th Judicial Circuit could lead to sanctions that include, “contempt; striking of pleadings or dismissal of action; fines and/or the imposition of attorney’s fees; and referral to The Florida Bar for disciplinary proceedings.”

Article: Why Attorneys Can Ethically Use General-Purpose GenAI for Client Matters Without Redacting Everything

 Carolyn Elefant's new article Why Attorneys Can Ethically Use General-Purpose GenAI for Client Matters Without Redacting Everything is available here.

Federal judge imposes what may be the most expensive sanction for hallucinations yet

 Last month I reported on a case that was reported as imposing significant sanctions for citing hallucinated cases and stated that the running count of cases involving courts complaining about hallucinated cases created by AI was 768.  Today the count is 901.  Go here for the full list.

I have not read every case on the list nor have I followed it closely, but if the case I reported in March was really significant because the court imposed $15,000 in sanctions to the lawyers involved, you should know that a federal judge has raised the stakes considerably.

Today, the ABA Journal is reporting on a case in which the judge imposed $110,000 in fines and attorney fees against two lawyers who filed documents filled with fake cases and fabricated citations.

That may be the larges sanctions imposed for improper use of AI yet.  

The ABA Journal has the story here.

Sixth Circuit imposes significant sanctions for citing non-existent cases

Last January I reported on a case imposing significant sanctions for citing hallucinated cases and stated that the running count of cases involving courts complaining about hallucinated cases (created by AI) was 527.  Two months later, the number is now 768 (and by the time you read this, it may be higher.)  Go here for the full list.

I am writing about this again today because LawSites is reporting (here) that "[i]n what may be one of the most significant appellate sanctions rulings yet involving fabricated case citations, the 6th U.S. Circuit Court of Appeals has imposed substantial penalties on two Tennessee attorneys for filing briefs containing more than two dozen fake or misrepresented citations."  

The lawyers involved were ordered to pay $15,000 each in fines and to jointly pay for their opponents' full attorney fees on appeal plus and amount equal to double their costs.

Yet another case on improper use of AI in litigation

 The running count of cases involving courts complaining about hallucinated cases (created by AI) is now at 527.  (See here.)

Based on a recent one, Stacie Rosenzweig (in her post "We Should All Know Better By Now, But We Don't") reminds us of some basic rules of thumb:  

If you’re going to use AI to generate arguments or citations, verify every single one of them with a trusted legal research source.

Second, if despite your best efforts something sneaks through, fess up as soon as you learn about it, and move to correct or withdraw the pleading. Chances are, the opposing attorney won’t object.

Third, if you’re a supervisor, get a good AI policy in place before things go sideways, and train your junior attorneys and staff. And, if you’re a supervisor on a specific matter, . . . it may be time to trust less, and verify more.

And, it’s definitely time to trust less, verify more if your subordinate has already shown they’ve had problems with misuse of AI. . . . 

Finally—it’s not just this case. I’m seeing more of a duty of opposing counsel to detect, and to report, bad citations (AI or otherwise) rather than wait for the court to find them out. It’s always been a best practice to read each case cited by the opposing party (at least, each case substantively cited) so you can adequately respond to their arguments; finding out that a case doesn’t exist or that it’s not at all reflective of why it was cited shouldn’t be too much of a lift.

By the way, the sanctions imposed in the case were harsh:

The court declined to impose monetary sanctions. Instead, all three plaintiff attorneys were disqualified (and the whole case was stayed to allow the plaintiff to find new counsel), and they were ordered to provide copies of the order to the presiding judge in every pending case in which they were counsel of record (and the clerk was directed to send a copy of the order to the Mississippi regulatory authorities). The firm was directed to perform an audit, and, in the paragraph that made me shudder the hardest, the associate was ordered to seek withdrawal from every case in which she was appearing before that judge, and was forbidden from appearing in any other case before that judge for a period of two years.

And then think about what these sanctions meant for the clients and the firm.  Clients in unrelated cases lost their choice of counsel and the firm was forced to explain to other clients why that was the case.  I wonder if the clients will take their business elsewhere and whether the attorney will be seeking a new job... 

ABA Task Force Report on AI in the practice of law

 I recently attended a program on the "Top Ten" topics in legal ethics for 2025 and I bet you will not be surprised to learn that number one was a collection of issues related to the use of AI in the practice of law.  And, for that reason, it is important to report that the ABA Task Force on Law and Artificial Intelligence recently published its 2025 report.

As reported by Bob Ambrogi in Law Sites, 

Artificial intelligence is no longer an abstract or experimental technology for lawyers – it is rapidly becoming core infrastructure for law practice, courts, legal education and access-to-justice efforts, and the legal profession must now shift its focus from whether to use AI to how to govern, supervise and integrate it responsibly.

That is the central conclusion of a report released yesterday from the American Bar Association’s Task Force on Law and Artificial Intelligence, a 56-page assessment that takes stock of how AI is already reshaping the profession and discusses the risks, opportunities and unresolved challenges that lie ahead.

The report, Addressing the Legal Challenges of AI: Year 2 Report on the Impact of AI on the Practice of Law, arrives at what the Task Force calls a “pivotal moment” for the profession. AI adoption has accelerated dramatically over the past year, pushing lawyers, judges, regulators and educators into unfamiliar terrain that demands new ethical frameworks, governance models and competencies.

You can get a free copy of the report here. 

Look out for more commentary in the next few weeks.  I just saw the first one in Above the Law here.

Illinois Bar Guide for Lawyers on Implementing AI Responsibly

 The Illinois Attorney Registration and Disciplinary Commission (ARDC) recently released “The Illinois Attorney’s Guide to Implementing AI” (the “Guide”), providing a practical framework for legal professionals to understand and deploy AI in their practices while honoring the ethical demands of confidentiality, data security, and supervision.

For more on the story go here.

IAALS evaluates Utah's "regulatory sandbox"

In 2020, the Utah Supreme Court launched the nation’s first legal regulatory sandbox, arguably designed to open the door to new kinds of legal service providers and business structures and with the overall goal of expanding access to legal services while still protecting consumers.

Five years later, the Institute for the Advancement of the Americal Legal System (IAALS) is starting to publish its findings as an independent evaluator of the Sandbox.

The current version of their report is being published across a four-part report series.  The reports will be released on a rolling basis.  The first two are now available here, and it looks like the next two will also be posted to this same page, so bookmark it if this is something you are interested in.

How not to practice law: cite hallucinated cases in proceeding to determine if you are going to be readmitted to practice

 In another of installment of dumb things that lawyers do, today we have the story of a suspended lawyer  who was caught citing at least one hallucinated case likely generated by artificial intelligence in his bid to return to law practice.  You would think that's a sure way to derail your own case, by showing how unfit you are to practice!  The ABA Journal has the story here.

Is there a duty to report another lawyer's use of fake, AI generated citations?

 As I am sure you know, Model Rule 8.3(a), which has been adopted in some form or another in every state, holds that "[a] lawyer who knows that another lawyer has committed a violation of the Rules of Professional Conduct that raises a substantial question as to that lawyer's honesty, trustworthiness or fitness as a lawyer in other respects, shall inform the appropriate professional authority."

So what happens when a lawyer notices that their opponent's court documents contain fake-"halucinated"-AI generated quotes or citations?  Does the lawyer have a duty to report it to the disciplinary authorities?  Evidently, it would be tactically wise to report it to the court in the litigation, and to use it against the opposition in argument, but the issue is not whether to report it to the court, the issue is whether to report it to the disciplinary authorities.

To my knowledge this question has not been addressed but a strict reading of the text of the rule suggests there is a duty to report, and a recent case in California might provide some support for that conclusion.

As reported in LawSites, 

A new decision from the California Court of Appeals adds an intriguing dimension to the growing body of AI hallucination sanctions cases, raising the question of a lawyer’s duty to detect fabricated, AI-generated citations — not in the lawyer’s own filings, but in an opponent’s.

While the court did impose a $10,000 sanction on the attorney who filed two appellate briefs containing fake citations, it also declined to award attorneys’ fees or costs to the opposing counsel, because of counsel’s failure to report the fake citations to the court or even to detect them.

That makes this what may be the first judicial decision to touch on on whether lawyers have a duty to detect and report their opponents’ AI-generated fake citations.

Of course, Rule 8.3 is based on "knowledge" and a lawyer could say that they did not know the opponents cases were fake, but then it could be argued that admitting to that is admitting to violating the duty of competence or diligence.  

So, it is possible that not only there is a duty to report, but that there is a duty to investigate and determine the validity of the opponent's sources in order to determine if the duty to report applies. 

Court Kicks Lawyers Off Case After Finding Fake AI Cases In Filings

 As more and more cases of fake citations, and faulty use of AI in lawyering pile up, judges may be ready to start imposing stronger sanctions.  Read the story here.

Yet another case of sanctions for relying on AI to write briefs (resulting in cites to cases that do not exist)

Two attorneys who were representing MyPillow CEO Mike Lindell in a defamation case in Denver have been fined $3,000 each as sanctions for submitting an inaccurate, AI-generated brief to the court in April.  Go here for the story.  

Above the Law has more on the story here, arguing the sanction was too low.

NPR has a comment here arguing that the case serves as a stark warning.

How not to practice law: AI edition

After a long absence from the blog today we have the second day in a row with a "how not to practice law' entry!  Today's story features a lawyer using AI hallucinated cases in support of a claim who, when called out on it, then cited more hallucinated cases.  That is bad enough, but the article also focuses on the fact that the trial judge issued an order based on the hallucinated cases... You can read the full story here,

Comment on Puerto Rico's newly adopted rule on technological competence

 As you probably know, paragraph 8 of the comment to Model Rule 1.1 (Competence) states, among other things, that "to maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology."  This statement has since been adopted and integrated into the comments of the rules in about 40 states.

Puerto Rico, however, recently adopted new rules (based on a translation of the Model Rules) and the notion of technological competence appears in a specific rule rather than in a comment to the rule on competence.  

In a comment posted in Above the Law, the author argues that "the Puerto Rico Supreme Court came out with a much more realistic and impactful definition of a lawyer’s technological competence obligations than that promulgated by some 40 states" and that "the Puerto Rican rules are a stronger statement about competence and suggest why it is so important."

You can read the full comment here.

A lawyer’s duties in response to a data breach

 Over at Ethical Grounds, the unofficial blog of Vermont's bar counsel, Michael Kennedy reminds us of a lawyer’s duties in response to a data breach.

DoNotPay now has to pay because of its claims about its non-existent AI lawyer

 Do you remember last year's stories about "DoNotPay," a company that claimed to have the "world's first robot lawyer"?   In case you don't, this was a company that claimed to offer legal services using a "robot lawyer" (aka "AI technology).  It made a lot of claims about the services it could provide, only it was revealed to be all smoke and mirrors.  I reported on the story throughout 2023 on  January 29, February 14, February 16, March 4, March 10, and March 17.

The story has made it back to the headlines of the week because on Wednesday, the FTC announced actions taken against five different companies making misleading claims about AI, including DoNotPay.  

In reply, DoNotPay as agreed to pay $193,000 to settle Federal Trade Commission allegations.  Besides paying the $193,000, DoNotPay agreed to refrain from making claims about its ability to substitute for legal services without evidence to support them.  

TechDirt has more details on this latest chapter on the DoNotPay saga here.  The ABA Journal has more on the settlement here.

Breaking Down the ABA’s Guidance on Using Generative AI in Legal Practice

A recent article breaking Down the ABA’s guidance on using generative AI in legal practice is available here.

Emerging Trends in Legal Tech

 The Legal Talk Network has a new article with links to podcasts on emerging trends in legal tech here.