Table 42

ai-sanctions-wave-part2

Posted on by root

AI Sanctions Wave – Part 2: Oregon’s Arithmetic

One state has turned vague judicial deterrence into calculable risk. Now other jurisdictions are asking whether to follow Oregon’s formula.

—

When Gabriel A. Watson appeared before the Oregon Court of Appeals in December 2025, the court did something unusual: instead of issuing a vague sanction based on judicial discretion, judges applied a mathematical formula. Two fake citations at $500 each, plus one fabricated quotation at $1,000: $2,000, exactly.[1]

That formula—$500 per fabricated citation, $1,000 per fabricated quotation—has become the cornerstone of Oregon’s approach to AI-related sanctions. Adopted by both the Oregon Court of Appeals and the U.S. District Court for the District of Oregon, it transforms the unpredictable threat of judicial displeasure into risk that law firm billing calculators can compute.

Three months later, Oregon alone accounted for $109,700 of the $145,000 nationwide Q1 2026 total—the largest jurisdictional aggregate for AI-related penalties in United States history.[2] The state’s contribution to the first quarter’s enforcement wave is no accident: Oregon established the formula first, applied it consistently, and now faces the question of whether its arithmetic represents a model for other jurisdictions.

This installment breaks down the key Oregon cases, analyzes how the formula works, and examines whether the state’s approach will spread.

—

The Formula Breakdown

The Oregon court didn’t arrive at its arithmetic by coincidence. The rationale was implicit: predictable penalties require predictable standards.[3]

The Formula: – $500 for each fabricated citation – $1,000 for each fabricated quotation
Why citations and quotations? – Citations: References to legal authorities that don’t exist or misrepresent holdings – Quotations: Fabricated quotes attributed to real or imaginary cases—a more serious violation because they create the appearance of direct authority
Formula Application Examples:

– Watson (December 2025): 2 citations × $500 + 1 quotation × $1,000 = $2,000[4] – Ghiorso (March 2026): 15 citations × $500 + 9 quotations × $1,000 = $16,500 minimum, capped at $10,000[5] – Couvrette v. Wisnovsky (March 2026): 15 citations × $500 + 8 quotations × $1,000 = $15,500 plus adverse costs[6]

The U.S. District Court for the District of Oregon adopted the same formula from the state appellate court, creating a unified sanctions landscape across state and federal courts within Oregon’s borders. Lawyers practicing in Oregon now know exactly what a hallucination will cost.

—

Case Study 1: Gabriel A. Watson — The Precedent

Attorney: Gabriel A. Watson, Portland-based attorney Court: Oregon Court of Appeals Date: December 2025 Infractions: 2 fabricated citations + 1 fabricated quotation Penalty: $2,000[7]

Watson’s case was the first to establish Oregon’s per-infraction formula. The court’s opinion identified AI-generated fabrications in Watson’s appellate filing and, rather than imposing an arbitrary fine, articulated a clear methodology: $500 for each fake citation, $1,000 for each fake quotation.

Why these numbers? The reasoning is implicit: citations are serious violations, but quotations—fabricated quotes attributed to legal authority—are worse. A fake citation alone suggests inadequate verification. A fake quotation suggests deliberate deception.

The Watson case carried precedent-setting weight. When courts outside Oregon began imposing AI-related sanctions in 2026, lawyers cited the Oregon formula in their own jurisdictions, urging judges to adopt similar calculable approaches.

—

Case Study 2: Couvrette v. Wisnovsky — The Federal Adoption

Court: U.S. District Court for the District of Oregon Practice Area: Intrafamily winery dispute Infractions: 15 fabricated citations + 8 fabricated quotations Penalty: $15,500 plus adverse costs[8]
Couvrette v. Wisnovsky was significant for two reasons. First, the federal district court adopted the Oregon Court of Appeals’ formula—demonstrating cross-jurisdictional consistency. Second, the case raised questions about client versus attorney responsibility.

The procedural timeline was revealing: Three summary judgment briefs were filed over five months, each containing AI-generated fabrications. The sanctions litigation that followed revealed evidence suggesting the client may have participated in generating the hallucinated citations.[9]

If the client used AI directly, was the attorney still responsible? The court answered yes: lawyers have a supervisory duty under Model Rule 5.3, regardless of who actually ran the chatbot. The $15,500 penalty—the full application of the formula—plus adverse costs made clear that attorneys bear responsibility for everything submitted under their name.

The federal adoption of Oregon’s formula created a unified landscape: whether lawyers appeared in state or federal court in Oregon, the same arithmetic applied.

—

Case Study 3: William Ghiorso — The Cap

Attorney: William Ghiorso, Salem-based civil attorney Court: Oregon Court of Appeals Date: March 2026 Infractions: “At least 15 fabricated citations and 9 quotations contrived from thin air”[10] Formula Minimum: $16,500 (15 × $500 + 9 × $1,000) Actual Penalty: $10,000 (capped)

Ghiorso’s case revealed a tension in the formula: when the number of infractions is large, the math can produce penalties that seem disproportionate. The court faced a question: apply the formula strictly ($16,500), or show leniency?

The court chose leniency, capping the penalty at $10,000. The rationale: Ghiorso had adopted new verification procedures and was responsive to remediation efforts. The formula allowed for flexibility; the court exercised it.[11]

But the cap raised questions about the formula’s consistency. If another attorney committed identical violations a month later, would the same cap apply? Or had the court created a ceiling that future litigants could cite? The uncertainty didn’t undermine the formula’s general application—Ghiorso still faced a substantial penalty—but it introduced the first judicial discretion challenge to Oregon’s arithmetic.

Public Coverage:

The Ghiorso decision attracted substantial media attention. OregonLive ran the headline “Oregon attorney slapped with record fine after citing case law hallucinated by AI.”[12] Gizmodo carried “Attorney Hit With Historic Fine.”[13] Salem Reporter provided detailed local coverage of the Salem attorney’s case.[14]

The narrative emerging in coverage: Oregon was becoming ground zero for AI sanctions enforcement. When other states were still deciding how to handle hallucinations, Oregon had a formula—and was applying it.

—

The Aggregate: $109,700

The $109,700 Oregon Q1 2026 total includes:

– Watson ($2,000)[15] – Couvrette v. Wisnovsky ($15,500 + adverse costs)[16] – Ghiorso ($10,000)[17] – Additional state-level sanctions (estimated $82,000+ from other cases)[18] – Costs and attorney fees across multiple matters

The aggregate matters for three reasons:

First, single-jurisdiction leadership: No other state even approaches Oregon’s totals. This isn’t because Oregon lawyers use AI more than lawyers elsewhere; it’s because Oregon courts established clear standards and applied them consistently. When other jurisdictions were issuing warnings or case-by-case sanctions, Oregon created a predictable enforcement regime.
Second, systematic enforcement: The aggregate suggests pattern enforcement, not isolated cases. Multiple judges across multiple courts in Oregon—all applying the same formula—signal a coordinated approach to AI-related misconduct. Lawyers practicing in Oregon now know that AI hallucinations will be discovered and penalized.
Third, external pressure: Other jurisdictions watching Oregon’s approach face a decision: follow the formula, or develop their own systems. The pressure comes not from federal mandate, but from law practice itself: lawyers want consistency across state lines. If Oregon’s formula works, other states will adopt it. If it fails, they’ll chart different paths.

—

The Deterrence Calculus

Transforming sanctions from judicial discretion to mathematical formula creates an unexpected benefit: lawyers can now run cost-benefit analyses on AI shortcuts. The deterrence calculus becomes explicit.

Traditional Westlaw/Lexis Research: – Cost: $2-5 per search (varies by firm contract) – Time: 30-120 minutes per brief – Citation verification: Built into tool (database-driven, pre-verified) – Risk: Virtually no sanction risk (tools are reliable)
AI-Assisted Research: – Cost: $20-100 per brief (flat-rate subscriptions or pay-per-query) – Time: 10-30 minutes per brief (AI speeds drafting) – Citation verification: Additional step if not built into workflow – Risk: Calculable sanction risk under Oregon’s formula
Sanction Risk Calculation:

Consider a brief with 20 citations and 3 quotations:

– If 5 citations are fake: 5 × $500 = $2,500 – If 2 quotations are fake: 2 × $1,000 = $2,000 – Total risk: $4,500 if verification fails

Compare that to the litigator’s economics:

– Typical billable hour: $400-800 – Brief drafting time: 2-6 hours = $800-4,800 of potential billable time – Sanction equivalent: 3-11 billable hours lost

The question is no longer vague: “Will the judge be upset?” It’s now arithmetic: “Is risking $4,500 worth saving 30 minutes of research?”

Law firm risk calculators: According to legal technology industry reports, multiple large law firms have begun deploying internal risk calculators that factor jurisdiction-specific formulas into associate training. Associates in Portland know the math: verify every citation, or $500 per error. Associates in other states have no comparable guidance—yet.

Not all lawyers respond to arithmetic the same way. Some will view $4,500 risk as acceptable cost-benefit tradeoff. But many, especially younger associates with eyes on partnership, will view any sanction risk as disqualifying.

—

Do Other Jurisdictions Follow?

Adoption Patterns:

– Oregon: State court and federal court using consistent formula; $109,700 Q1 2026 aggregate – California: No formula established yet; case-by-case sanctions based on judicial discretion – New York: Following Mata v. Avianca precedent; stricter than many states but no per-infraction formula – Texas: Early sanctions cases in 2025-26; no consistent formula emerging – Federal Circuits: Sixth Circuit substantial penalty ($30,000 in Whiting v. Athens); other circuits observing but not adopting similar approaches

Resistance Factors:

Several factors explain why jurisdictions resist adopting Oregon’s formula:

1. Judicial preference for discretion: Many judges believe sanctions should be tailored to specific circumstances—prior disciplinary history, severity of misconduct, remediation efforts. A formula removes judicial discretion; some courts value that flexibility.

2. Concern that formula may encourage corner-cutting: If lawyers know the exact cost of a mistake, some may deliberately skip verification when they can afford the penalty. The Oregon formula transforms sanctions from unknown risk to calculable cost.

3. State bar association variation: Ethical standards vary by state. What qualifies as “reasonably thorough” verification under Oregon ethics rules may differ from California or New York interpretations. A one-size-fits-all formula may not align with local practice norms.

Evidence of Diffusion:

Despite resistance, evidence suggests Oregon’s formula is influencing other jurisdictions:

– California: Defense attorneys have begun citing the $500/$1,000 formula in motions arguing for calculable sanctions standards. – New York: The 2023 Mata v. Avianca opinion is frequently referenced alongside Oregon cases, creating de facto pressure toward consistent penalties. – Federal circuits: The Sixth Circuit’s $30,000 penalty in Whiting v. Athens—while not adopting Colorado’s formula—reflects an escalation trend that Oregon率先开创served as the early warning.

Oregon isn’t imposing its formula on anyone. But by going first—the first state to establish clear, predictable AI-related sanctions—Oregon created a reference point. Other jurisdictions can now compare their own approaches to Oregon’s arithmetic and ask: Is our system working better?

—

Is the Formula Working?

The evidence is mixed.

Arguments for Effectiveness:

– Oregon cases in Q1 2026: While the state led in total penalties ($109,700), Charlotin’s database shows acceleration in other states—suggesting Oregon’s consistency may have created deterrence within its borders that other jurisdictions lack.[19] – Legal tech vendor response: Some vendors have begun advertising “verification features” specifically targeted at Oregon practitioners (“Citations with 100% accuracy—no Oregon-style sanctions”). Product development responds to perceived demand; if Oregon lawyers feared sanctions, vendors built features to address that fear.

Arguments Against Effectiveness:

– Q1 2026 saw new Oregon cases: Despite the formula being established in December 2025, attorneys continued submitting hallucinated citations throughout Q1 2026—including William Ghiorso’s record-setting case. The formula didn’t prevent violations; it only quantified penalties. – Pro se litigants unaffected: The formula applies to lawyers, not self-represented litigants with no bar licenses to suspend. Pro se parties continue appearing in Charlotin’s database, unaffected by arithmetic-based deterrence. – AI tool use continues: Legal tech industry reports show no decline in AI adoption rates in Oregon. Lawyers continue using generative AI tools; the formula changed verification practices, not tool selection.

The formula is a sanctions mechanism, not a prevention system. It provides clarity about consequences, not technical solutions to hallucination risk. Lawyers who want to use AI safely still need verification processes that Oregon’s arithmetic alone doesn’t provide.

—

What’s Next for Oregon?

The state faces three key questions in late 2026 and beyond:

1. Will other states adopt the formula?

The pressure is growing for interstate consistency. Lawyers litigating across state lines face wildly varying sanctions regimes—$500 per error in Oregon, judicial discretion in California, escalating federal penalties in appellate courts. Bar associations and state supreme courts will face calls to standardize. Oregon’s formula provides a concrete proposal that can be adapted, adopted, or rejected.

2. Will the formula evolve?

Ghiorso’s case revealed a tension: large numbers of infractions can trigger disproportionate penalties under strict formula application. Oregon courts may need to clarify when caps apply, whether there’s a maximum penalty threshold, and how remediation efforts should factor into arithmetic-based sentencing.

3. Will the formula address non-citation misconduct?

AI generates more than just fake citations. Drafted arguments can contain hallucinated factual claims. Summary judgment briefs can misquote deposition excerpts. Oregon’s $500/$1,000 framework addresses citations and quotations—but what about paragraph-length fabrication? The state will need to decide whether to expand the formula’s scope or develop secondary mechanisms.

—

The Broader Question

Oregon’s arithmetic represents a broader choice: when technology creates new forms of professional misconduct, how should legal systems respond?

Two paths are possible:

Path 1: Discretionary enforcement (traditional approach) – Treat each case individually – Judges tailor penalties to circumstances – Flexibility but unpredictability – Lawyers lack clear deterrence signals
Path 2: Formula-based enforcement (Oregon’s approach) – Fixed penalties for defined violations – Predictable calculus but rigid application – Risk calculators and cost-benefit analysis – Clear deterrence signals but potential exploitation

Oregon chose Path 2. The first quarter of 2026 provided real-world data on how that choice plays out. Other jurisdictions are watching—some with interest, some with skepticism, and all with the knowledge that AI-related cases are not going away.

The question isn’t whether Oregon’s formula is perfect. The question is whether it’s better than the alternative in an age when lawyers must choose between traditional research methods and AI-assisted shortcuts. Oregon answered: calculable risk beats unknown consequence.

Next: The Sixth Circuit Line

Part 3 examines Whiting v. City of Athens, a March 2026 federal appellate decision that imposed $30,000 in direct fines and articulated a tool-agnostic principle that may reshape enforcement across the federal judiciary. —

Citations

—

Series Navigation

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

Notes

  1. *Ringo v. Colquhoun Design Studio, LLC*, 345 Or. App. 301 (December 2025), https://law.justia.com/cases/oregon/court-of-appeals/2025/a186670.html (per-infraction formula established: $500/citation, $1,000/quotation); official opinion PDF: https://websitedc.s3.amazonaws.com/documents/Ringo_v._Colquhoun_Design_Studio_USA_3_December_2025.pdf ↩

  2. 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/ ↩

  3. *Ringo v. Colquhoun Design Studio, LLC*, note 1 (“predictable penalties require predictable standards”) ↩

  4. *Ringo v. Colquhoun Design Studio, LLC*, note 1 ↩

  5. 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 ↩

  6. *Couvrette v. Wisnovsky*, No. 3:24-cv-01444-SI (D. Or. Feb. 27, 2026); NWSidebar (WSBA), “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/ ↩

  7. *Ringo v. Colquhoun Design Studio, LLC*, note 1 ↩

  8. *Couvrette v. Wisnovsky*, note 6 ↩

  9. *Couvrette v. Wisnovsky*, note 6 (court noted client Joanne Couvrette may have collaborated in generating fabricated citations); see also San Diego Union-Tribune, “San Diego attorney hit with one of largest ever sanctions for submitting AI-hallucinated filings,” April 4, 2026, https://www.sandiegouniontribune.com/2026/04/04/san-diego-attorney-hit-with-one-of-largest-ever-sanctions-for-submitting-ai-hallucinated-filings/ ↩

  10. OregonLive, note 5; 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 ↩

  11. Oregon Court of Appeals, *Ghiorso v. [defendant]*, March 2026 ($10,000 fine, 15 fabricated citations, 9 fabricated quotes, capped at $10,000); see 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 ↩

  12. OregonLive, note 5 ↩

  13. 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 ↩

  14. Inc., note 10 ↩

  15. *Ringo v. Colquhoun Design Studio, LLC*, note 1 ↩

  16. *Couvrette v. Wisnovsky*, note 6 (claims dismissed with prejudice; lead attorney Stephen Brigandi fined $15,500; local counsel Peter Murphy fined $14,000 for failure to supervise) ↩

  17. *Ghiorso*, note 11 ($10,000 cap applied despite formula yielding higher amount) ↩

  18. EDRM/ComplexDiscovery analysis, note 2 (estimated additional Oregon cases contributing to $109,700 aggregate) ↩

  19. Damien Charlotin, AI Hallucination Cases Database, Sciences Po Law School & HEC Paris Smart Law Hub, https://www.damiencharlotin.com/hallucinations/; CSV download: https://www.damiencharlotin.com/hallucinations/hallucinations/download.csv ↩

  20. Eugene Volokh, “$14K Sanction for Local Counsel’s Not Meaningfully Supervising Out-of-Jurisdiction Counsel,” *Reason (Volokh Conspiracy)*, March 24, 2026, https://reason.com/volokh/2026/03/24/14k-sanction-for-local-counsels-not-meaningfully-supervising-out-of-jurisdiction-counsel/ ↩

  21. Model Rules of Professional Conduct, Rule 5.3 (Supervision), https://www.americanbar.org/groups/professional_responsibility/resources/lawyer_ethics_regulation/model_rules_of_professional_conduct/rule_5_3_responsibilities_regarding_nonlawyer_assistants/ ↩

  22. 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 ↩

  23. 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/ ↩

  24. Portland Business Journal, “An Oregon judge slaps a fine on a lawyer for ‘notorious’ use of AI in briefs,” December 2025, https://www.bizjournals.com/portland/news/2025/12/17/wisnovsky-lawsuit-generative-ai-lawyer-judge.html ↩

  25. KGW, “Oregon lawyer fined $10,000 for using false AI info in legal brief,” March 2026, https://www.kgw.com/video/news/local/oregon-lawyer-fined-10000-for-using-false-ai-info-in-legal-brief/283-09c38c5a-259b-4fca-a10c-90a6393f2f53 ↩

  26. Noah News, “Courts escalate sanctions as AI hallucinations in legal filings surge in 2026,” 2026, https://noah-news.com/courts-escalate-sanctions-as-ai-hallucinations-in-legal-filings-surge-in-2026/ ↩

  27. Federal Rules of Civil Procedure, Rule 11, https://www.law.cornell.edu/rules/frcp/rule_11 ↩

  28. ABA 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/ ↩

  29. *Couvrette v. Wisnovsky*, note 6 (23 hallucinated citations across 3 documents) ↩

  30. *Ghiorso*, note 11 (15 fabricated citations, 9 fabricated quotes, “contrived from thin air”)

    — ↩