Table 42

ai-sanctions-wave-part1

Posted on by root

AI Sanctions Wave – Part 1: The $145,000 Quarter

When lawyers started asking ChatGPT to write their briefs, courts sent them home with warnings. Three years later, judges have lost patience.

—

In the first three months of 2026, U.S. courts imposed at least $145,000 in sanctions for lawyers who filed documents containing AI-generated fake citations.[1] That figure—a single quarter’s worth of penalties—dwarfs every prior AI-related sanction on record and marks a turning point in how the American legal system is responding to generative artificial intelligence.

January 2026 saw $5,000 in nationwide AI-related sanctions. February, just $250. Then March exploded: $139,750 in penalties arrived within weeks, including a single federal appellate case that alone cost two attorneys $30,000 in direct fines before attorney fees and double costs.[2]

What happened between February and March? Judicial patience, it appears, finally ran out.

This six-part series examines the architectural implications of this sanctions wave for legal practice: What standards are emerging? Who bears responsibility for AI-generated content—lawyers, judges, or developers? Will these penalties reshape how attorneys work, or represent a temporary adjustment period as the profession adapts to new technology?

The first quarter of 2026, we’ll see, is not just a statistical anomaly. It’s the moment courts moved from treating AI-related filing failures as surprising novelty to enforcing them as professional misconduct.

—

The Data: From Strange Exception to Enforceable Pattern

The $145,000 Q1 2026 total comes from court docket tracking compiled by Damien Charlotin, a legal technology researcher at HEC Paris’s Smart Law Hub. Charlotin’s AI Hallucination Cases Database has documented AI-related filings since 2023, and his findings on the first quarter of 2026 reveal an unmistakable acceleration.[3]

The monthly breakdown tells the story:

– January 2026: $5,000 in sanctions – February 2026: $250 in sanctions – March 2026: $139,750 in sanctions

Oregon alone accounts for $109,700 of the Q1 total—the largest jurisdictional aggregate in the United States.[4] The remainder comes primarily from the Sixth Circuit’s federal appellate decision in Whiting v. City of Athens ($30,000 direct fines) and additional state-level penalties scattered across the calendar.[5]

But the aggregate figure understates the acceleration. Charlotin’s database shows a surge in early March: at one point, ten separate courts flagged AI-generated fake citations on a single day. “The pace has reached relentless levels,” Charlotin told NPR.[6]

What changed? Courts didn’t announce new policies in March. No bar associations suddenly issued emergency rules. The acceleration, instead, represents the cumulative effect of three years of educational approaches proving insufficient.

Update (April 2026): The Q1 total of $145,000, while significant, has already been eclipsed. By early April, the largest single AI sanctions award on record reached $86,000 in ByoPlanet v. Johansson (S.D. Fla.), involving repeated systemic misuse across multiple filings despite prior warnings. The Sixth Circuit issued a second published AI sanctions opinion in United States v. Farris (April 3, 2026), and the Fifth Circuit entered the enforcement picture with its published opinion in Fletcher v. Experian (February 2026). Aggregate tracking now documents 1,031+ cases globally and 128+ individual lawyers implicated in the United States, with Am Law 100 firms among the sanctioned. The Q1 wave, it turns out, was not a peak—it was a foothill.[^update1]

—

Why Q1 2026? The Perfect Storm

Several factors converged to make the first quarter of 2026 the moment enforcement accelerated:

Three Years of Learning Curve: It has been more than three years since Mata v. Avianca, the 2023 New York case that made AI hallucinations a national punchline. In that case, federal district Judge P. Kevin Castel sanctioned an attorney when ChatGPT generated entirely fabricated case citations—complete with imaginary quotes and docket numbers.[7] Initially, courts took an educational approach: warnings, orders to show cause, requests for explanations. By early 2026, judges had seen enough. The novelty had worn off; tolerance had not.
Professional Guidance Convergence: The American Bar Association issued Formal Opinion 512 in July 2024, establishing the first comprehensive framework for AI use in legal practice under the Model Rules of Professional Conduct.[8] By Q1 2026, state bar associations had begun issuing supplemental guidance. The rules were now clear enough that ignorance was no longer a credible defense.
Tool Proliferation: Commercial legal AI tools—Westlaw CoCounsel, Lexis+ AI, Perplexity—became widely available in late 2025. Subscription-based generative AI lowered the barrier to entry: lawyers could now access sophisticated legal research assistants for flat monthly fees. More tools meant more AI-generated filings—and more opportunities for hallucinated citations to slip through.[9]
Judicial Patience Exhausted: Pattern recognition matters. Judges who saw one or two AI hallucination cases in 2024 and 2025 now see them regularly. What was once a strange exception requiring explanation became a predictable pattern demanding enforcement. The Sixth Circuit’s March 13, 2026 decision in Whiting v. City of Athens captured this shift explicitly: “Smaller fines had proven inadequate given what it described as an ongoing stream of cases presenting the same problems,” wrote Circuit Judge John K. Bush.[10]

The acceleration in Q1 2026 wasn’t driven by any single case or policy change. It was the point when enough judges had enough experience, enough cases, and enough frustration to start imposing meaningful penalties.

—

Who’s Getting Sanctioned?

Three categories dominate the Q1 2026 sanctions wave:

Lawyers: The most common category—attorneys using commercial legal AI tools or general-purpose chatbots for research and drafting. These cases typically involve higher sanctions, especially when attorneys have prior disciplinary records or when the scale of misconduct is substantial. In Whiting v. City of Athens, for example, attorneys Van R. Irion and Russ Egli faced $15,000 sanctions each after federal appellate judges found “over two dozen” fabricated citations across multiple appellate briefs. Both attorneys had prior disciplinary histories, and their refusal to comply with the court’s show-cause order aggravated the penalty.[11]
Pro Se Litigants: Representing themselves in court without attorneys, pro se litigants account for a large proportion of cases in Charlotin’s database.[12] Sanctions in these cases are typically smaller—often warnings or admonishments rather than substantial fines—but the pattern is clear: non-lawyers using AI without legal training face heightened risk of hallucinations.
Law Firms: Partner-level responsibility is emphasized in California and other jurisdictions under Model Rule 5.3, which holds supervising attorneys accountable for subordinates’ conduct. Some cases involve client participation in AI drafting, as in Couvrette v. Wisnovsky, an intrafamily winery dispute where the court suspected the client may have collaborated in generating the fabricated citations.[13]

The practice areas affected are equally varied: civil litigation (most common), family law disputes, bankruptcy proceedings, administrative hearings, contract disputes, and employment litigation. No area of practice appears immune when lawyers use AI shortcuts without verification.

—

The Sanctions Spectrum

The penalties imposed in Q1 2026 range from modest to career-threatening:

– Monetary fines: From $250 for minor infractions to $30,000+ for federal appellate misconduct[14] – Adverse costs: Requiring sanctioned parties to pay opposing counsel’s attorney fees[15] – CLE requirements: Mandatory continuing legal education sessions focused on AI and professional responsibility[16] – Brief striking: Entire motions or briefs removed from the record when contaminated with hallucinations[17] – Disqualification: Attorneys removed from cases when misconduct undermines their credibility[18] – Bar referrals: Disciplinary complaints to state bar associations for investigation[19] – Public admonishments: Judicial opinions explicitly naming and shaming offending counsel[20]

Oregon’s approach—establishing a predictable per-infraction fee schedule—has become particularly influential. In December 2025, the Oregon Court of Appeals established a formula in a now-cited precedent: $500 for each fabricated citation, $1,000 for each fabricated quotation.[21] That formula, adopted by both state and federal courts in Oregon, transforms vague deterrence into calculable risk.

The result: lawyers can now run the math. A brief with 20 citations and 3 quotations—if 5 of each are hallucinated—would cost exactly $5,500 in Oregon under the formula. Risk calculators have appeared on law firm intranets nationwide.

—

The Pattern Accelerates

The significance of Q1 2026 is not just the total figure, but the pattern it reveals. The acceleration from February’s $250 to March’s $139,750 is too dramatic to attribute to random variation. Something systemic shifted.

That shift, the data suggests, is that courts have moved from the “learning phase” to the “enforcement phase” of AI-related sanctions. Mata v. Avianca was 2023—three years is enough time for educational approaches to prove inadequate. The March 2026 explosion in penalties signals that judges are no longer treating AI-generated hallucinations as curious anomalies requiring explanation. They are treating them as professional misconduct requiring consequences.

The message to the profession is clear: the grace period is over. Courts expect lawyers to understand AI’s limitations, verify outputs, and face enforcement if they don’t.

Next: Oregon’s Arithmetic

Part 2 examines how Oregon emerged as ground zero for AI sanctions jurisprudence, analyzing the state’s per-infraction formula in detail and asking whether other jurisdictions will follow Oregon’s arithmetic. —

Citations

—

Series Navigation

– Part 1: The $145,000 Quarter (current) – Part 2: Oregon’s Arithmetic – Part 3: The Sixth Circuit Line – Part 4: The Judicial AI Paradox – Part 5: The Labeling Problem – Part 6: Upstream Liability

Notes

  1. EDRM/ComplexDiscovery, “The AI Sanction Wave: $145K in Q1 Penalties Signals Courts Have Lost Patience with GenAI Filing Failures,” April 6, 2026, https://complexdiscovery.com/the-ai-sanction-wave-145k-in-q1-penalties-signals-courts-have-lost-patience-with-genai-filing-failures/ ↩

  2. Damien Charlotin, AI Hallucination Cases Database, Sciences Po Law School & HEC Paris Smart Law Hub, https://www.damiencharlotin.com/hallucinations/ (accessed April 12, 2026) ↩

  3. Damien Charlotin, LinkedIn post, “The AI Hallucination Cases Database has reached [milestone],” March 4, 2026, https://www.linkedin.com/posts/damien-charlotin-45769278_the-ai-hallucination-cases-database-has-reached-activity-7434881585788760064-Mg6e ↩

  4. EDRM/ComplexDiscovery, note 1; Oregon aggregate confirmed in *Couvrette v. Wisnovsky*, No. 3:24-cv-01444-SI (D. Or. Feb. 27, 2026) ↩

  5. *Whiting v. City of Athens*, No. 25-5424 (6th Cir. March 13, 2026), https://law.justia.com/cases/federal/appellate-courts/ca6/25-5424/25-5424-2026-03-13.html ↩

  6. NPR, “Courts Impose $145,000 in Sanctions for AI-Generated Fake Citations in First Quarter Alone,” April 3, 2026 ↩

  7. *Mata v. Avianca, Inc.*, No. 22-cv-1461 (PKC) (S.D.N.Y. June 22, 2023), https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2022cv01461/575368/54/; See also Benjamin Weiser, “A Man Sued Avianca Airline. His Lawyer Used ChatGPT.,” *New York Times*, May 27, 2023, https://www.nytimes.com/2023/05/27/nyregion/avianca-airline-lawsuit-chatgpt.html ↩

  8. American Bar Association Standing Committee on Ethics and Professional Responsibility, Formal Opinion 512, “Generative Artificial Intelligence Tools and the Profession,” July 29, 2024, https://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/aba-formal-opinion-512/ ↩

  9. Thomson Reuters Institute, “Responsible AI Use for Courts,” January 2026, https://www.thomsonreuters.com/en-us/posts/wp-content/uploads/sites/20/2026/01/Hallucinations-Report-2026_FINAL.pdf (documenting tool proliferation and Charlotin database methodology) ↩

  10. *Whiting v. City of Athens*, 2026 WL 710568, at *8 (6th Cir. March 13, 2026) (Bush, J.) (“Smaller fines had proven inadequate given what it described as an ongoing stream of cases presenting the same problems”) ↩

  11. *Whiting v. City of Athens*, note 5; National Law Review, “Sixth Circuit Sanctions Attorneys for Fake Citations—What Does This Mean for Use of AI?,” March 14, 2026, https://natlawreview.com/article/sixth-circuit-sanictions-attorneys-fake-citations-what-does-mean-use-ai ↩

  12. Charlotin database, note 2 (pro se litigants constitute a large proportion of cases in the database; see download.csv at https://www.damiencharlotin.com/hallucinations/hallucinations/download.csv) ↩

  13. *Couvrette v. Wisnovsky*, No. 3:24-cv-01444-SI (D. Or. Feb. 27, 2026); See also NWSidebar (Washington State Bar Association), “Parade of Horribles: Federal Court in Oregon Surveys Sanctions for AI Fake Citations,” March 2, 2026, https://nwsidebar.wsba.org/2026/03/02/parade-of-horribles-federal-court-in-oregon-surveys-sanctions-for-ai-fake-citations/ ↩

  14. Sanctions range across multiple Q1 2026 cases documented in Charlotin database, note 2; see also Reuters, “US appeals court fines lawyers $30,000 in latest AI-related sanction,” March 16, 2026, https://www.reuters.com/legal/litigation/us-appeals-court-fines-lawyers-30000-latest-ai-related-sanction-2026-03-16/ ↩

  15. *Ghiorso v. [defendant]*, Oregon Court of Appeals, March 2026 (adverse costs awarded alongside $10,000 fine); OregonLive, “Oregon attorney slapped with record fine after citing case law hallucinated by AI,” March 2026, https://www.oregonlive.com/pacific-northwest-news/2026/03/oregon-attorney-slapped-with-record-fine-after-citing-case-law-hallucinated-by-ai.html ↩

  16. Various state court orders requiring AI-specific CLE, documented in Charlotin database, note 2; see also ABA Formal Opinion 512, note 8 (recommending competence requirements for AI use) ↩

  17. Multiple March 2026 cases in Charlotin database, note 2; see *Couvrette v. Wisnovsky*, note 13 (claims dismissed with prejudice) ↩

  18. Multiple Q1 2026 sanctions orders, Charlotin database, note 2; see also *Couvrette v. Wisnovsky*, note 13 (local counsel sanctioned $14,000 for failure to supervise) ↩

  19. *Whiting v. City of Athens*, note 5 (bar referrals ordered for both attorneys Van R. Irion and Russ Egli) ↩

  20. *Ringo v. Colquhoun Design Studio, LLC*, 345 Or. App. 301 (December 2025), https://law.justia.com/cases/oregon/court-of-appeals/2025/a186670.html; *Ghiorso*, note 15 ↩

  21. *Ringo v. Colquhoun Design Studio, LLC*, note 20 (per-infraction formula established: $500/citation, $1,000/quotation); adopted by federal court in *Couvrette v. Wisnovsky*, note 13 ↩

  22. Seyfarth Shaw LLP, “Update on the ChatGPT Case: Counsel Who Submitted Fake Cases Are Sanctioned,” June 2023, https://www.seyfarth.com/news-insights/update-on-the-chatgpt-case-counsel-who-submitted-fake-cases-are-sanctioned.html ↩

  23. Model Rules of Professional Conduct, Rule 11 (Federal Rules of Civil Procedure), https://www.law.cornell.edu/rules/frcp/rule_11 ↩

  24. Eugene Volokh, “Lawyers Citing Nonexistent Cases Ordered to Pay Opponents’ Attorney Fees, Double Costs, $15K Fine Each,” *Reason (Volokh Conspiracy)*, March 14, 2026, https://reason.com/volokh/2026/03/14/lawyers-citing-nonexistent-cases-ordered-to-pay-opponents-attorney-fees-double-costs-15k-fine/ ↩

  25. Bob Ambrogi, “Sixth Circuit Slaps Steep Sanctions on Two Lawyers for Fake Citations and Misrepresentations in Appellate Briefs,” *LawSites*, March 2026, https://www.lawnext.com/2026/03/sixth-circuit-slaps-steep-sanctions-on-two-lawyers-for-fake-citations-and-misrepresentations-in-appellate-briefs.html ↩

  26. *Mata v. Avianca, Inc.*, note 7 (Judge Castel’s sanctions order: $5,000 fine, letter to client, and copies to opposing counsel) ↩

  27. FindLaw, *Mata v. Avianca Inc* (2023), https://caselaw.findlaw.com/court/us-dis-crt-sd-new-yor/2335142.html ↩

  28. Paul Caron, “Worldwide Tally of Legal Decisions Involving AI Hallucinations,” *Tax Prof Blog (AALS)*, April 8, 2026, https://taxprofblog.aals.org/2026/04/08/worldwide-tally-of-legal-decisions-involving-ai-hallucinations/ ↩

  29. Gizmodo, “Attorney Hit With Historic Fine for Citing AI-Generated Cases,” March 2026, https://gizmodo.com/attorney-hit-with-historic-fine-for-citing-ai-generated-cases-2000738651 ↩

  30. Inc., “Faulty AI Leads to $10,000 Fine for Oregon Lawyer,” March 2026, https://www.inc.com/kevin-haynes/faulty-ai-leads-to-record-10000-fine-for-oregon-lawyer/91322007 ↩

  31. Northwestern University, “Federal Judges Report Broad Adoption of AI Tools,” March 30, 2026, https://news.northwestern.edu/stories/2026/03/northwestern-study-finds-a-significant-number-of-federal-judges-are-already-using-ai-tools ↩

  32. *Ringo v. Colquhoun Design Studio, LLC*, official opinion, December 2025, https://websitedc.s3.amazonaws.com/documents/Ringo_v._Colquhoun_Design_Studio_USA_3_December_2025.pdf ↩