Ad/Marketing Communication
This legal article/report forms part of my ongoing legal commentary on the use of artificial intelligence within the justice system. It supports my work in teaching, lecturing, and writing about AI and the law and is published to promote my practice. Not legal advice. Not Direct/Public Access. All instructions via clerks at Doughty Street Chambers. This legal article concerns AI Law.
Introduction
Thank you to everyone who has shared cases related to AI Hallucinations and the other three trackers I’m updating. All trackers are now available in one place [here].
For the AI Hallucination Cases Tracker, I try to invest significant time in carefully reading each judgment before adding my analysis. As a result, I may not always upload new cases as promptly as I’d like.
As I don’t practise in every jurisdiction, occasional inaccuracies might arise, though I strive for precision. While there are many excellent AI blogs available, I’ve recently been sent some that cause me concern. Their commentary on a case does not reflect my research and understanding, which could lead to misreported case details, reasoning, or outcomes. I encourage readers to always carefully review judgments themselves to ensure their understanding and conclusions are accurate.
I’m currently reviewing both new and existing entries to ensure they meet these standards. Though this process may take slightly longer than ideal, I hope the result justifies the wait.
If you notice inaccuracies or have relevant updates, I’d greatly appreciate your input. All contributions enhance the tracker’s accuracy and usefulness.
A Recent AI Hallucination Case to Add to the Tracker
In Mid Central Operating Engineers HWF v Hoosiervac LLC, Magistrate Judge Mark Dinsmeore recommended personal sanctions against a lawyer totalling $15,000:
“for submitting to the Court and opposing counsel, on three separate occasions, briefs that contained citations to non-existent cases”
The lawyer argued:
“any sanctions are “moot” because he has suffered “significant and irreversible harm to [his] professional reputation.”
The Judge stated some cases cited in support “…don’t apply here.” However, the Court acknowledged:
“That said, in considering an appropriate sanction the Court takes into account the steps [Lawyer] has taken “to educate himself on the responsible use of AI in legal practice” and adhere to “the highest standards of professional conduct moving forward.” See dkt. 102 at 2. The Court also considers the collateral consequences that [Lawyer] has experienced, and may continue to experience, from having improperly relied on non-existent AI-generated legal citations…so, the court has considered those circumstances alongside its interest in deterring careless or reckless attorney conduct.”
Ultimately, the Judge decided that the sanction should be reduced to $6,000, concluding:
“…suffices to deter repetition of the conduct or comparable conduct by others similarly situated.” Fed. R. Civ. P. 11(c)(4). This substantial sanction is appropriate because, as Magistrate Judge Dinsmore explained, [Lawyer] filed briefs containing nonexistent AI-generated legal citations on three separate occasions and the penalties imposed on other attorneys for similar misconduct “have evidently failed to act as a deterrent.” Dkt. 99 at 4–5 (collecting cases imposing sanctions up to $5,000). The Court therefore ADOPTS in part Magistrate Judge Dinsmore’s report and recommendation. Dkt. [99]. [lawyer] is personally sanctioned $6,000 and shall pay that amount to the Clerk of the Court by July 10, 20251.”
Commentary
This case continues the trend of courts taking a firm stance with robust sanctions against lawyers who rely on AI-generated “hallucinations.” However, it also highlights how judges weigh mitigating factors when assessing sanctions.
The lawyer argued that the incident caused “significant and irreversible harm to [his] professional reputation.” Undeniably, being publicly identified in judgments creates significant professional embarrassment, which, in my view, is a key mitigating factor when considering sanctions. For many, the public exposure alone will outweigh the financial penalty a court imposes, significantly deterring repeated misconduct.
However, I do not believe, as was suggested here, this reputational harm is entirely irreversible in many cases. Mistakes happen, even among experienced lawyers, that’s why we have professional insurance. Provided errors are swiftly acknowledged, openly addressed, and accompanied by genuine corrective actions, I believe reputational recovery is achievable.
That said, attempting to conceal errors or delaying acknowledgment until discovery makes reputational recovery far more challenging.
Other important mitigating factors, such as genuine efforts to educate oneself about responsible AI use, proactively referring to relevant oversight bodies and sincere apologies, go a long way towards mitigation. I’ll explore this in more detail in a future post.
What are your thoughts on this issue? Is reputational damage a sufficient deterrent? Can that reputational damage be repaired? The discussion continues on LinkedIn and my Substack and don’t forget to subscribe to my newsletter here for more insights and updates on emerging AI legal challenges.
