Adverse Costs Orders in AI Hallucination Cases: The 2026 Pattern

If you skim the Damien Charlotin AI Hallucination Cases database (~1,750 cases) and look for outlier penalty amounts, the pattern is not extraordinary direct fines. Direct fines cluster in the low four figures. The pattern is that adverse-costs orders — fee-shifting from the sanctioned side to the opposing side — are what scale the exposure. Every U.S. AI-hallucination sanction over $30,000 in 2026 involves an adverse-costs component. Every one over $80,000 is almost entirely adverse-costs, with a small direct-fine layer on top. This post walks the 2026 pattern and what it means for how firms should price the risk.

The direct-fine vs. costs-order split, illustrated

Couvrette v. Wisnovsky (D. Oregon, March 2026): $15,504 direct sanction + $94,700 adverse-costs order = $110,204 total. LiveVideo.AI Corp. v. Redstone (S.D.N.Y., June 2026): almost entirely the $80,056 adverse-costs order. Joel A. Rivera v. Triad Properties Corporation (N.D. Alabama, March 2026): $35,603, adverse-costs-dominant. Whiting v. City of Athens (6th Cir., March 2026): $30,000, adverse-costs order plus direct sanction. Laurie Ibach and Mark Stewart v. Bruce Stewart (SC Alabama, April 2026): $17,200, adverse-costs order plus doubled costs. The pattern: direct fines rarely clear five figures; adverse-costs orders routinely do.

Why direct fines have a soft ceiling

Rule 11 and equivalent state authority orient toward deterrence sanctions calibrated to the least severe amount that will deter the wrongful conduct. In practice, that anchors direct-fine amounts to what the court thinks will make an individual attorney pay attention — typically $500 to $10,000. Courts occasionally go higher, but the practical ceiling is meaningfully below what the same-case adverse-costs order can reach.

Why adverse-costs orders scale

Adverse-costs orders are pegged to the opposing side’s actual reasonable attorney fees on the sanctioned briefing. In a pro se case or a small-dollar dispute those fees are small. In a defended commercial case with senior partners on the other side, they are large — sometimes very large. The order shifts what the opposing side actually incurred. Couvrette’s $94,700 costs order is not a sanction the court thought up; it is a documented fee amount from the opposing party’s side of the case.

Small-dollar and pro se cases: the other end of the distribution

At the other end, small-dollar and pro se cases produce small numbers even when the sanction is imposed. Arrieta v. Cervantes (CA Arizona, July 2026): $1 nominal adverse-costs order. Villanueva v. Bowers (CA Arizona 1d, June 2026): $1 nominal. Nady v. Al Sadi (Ontario SCJ, June 2026): $1 CAD, but recorded as “AI misuse increasing adverse costs award” — a flag on the record with a nominal dollar amount. Reaves Law Firm v. Baker Donelson (W.D. Tennessee, June 2026): $1 nominal adverse-costs order plus bar referral. These orders are dollar-trivial and record-significant. They mark the case for downstream discipline and future risk pricing.

Nominal-dollar orders are not consolation prizes

The $1 orders in the Charlotin database are strategically important. They put a formal adverse finding on the attorney’s record without extracting money in a case where extraction would be inequitable — the sanctioned attorney is often unrepresented or the case has no fee shift to pool from. But the record entry is real, and it is available to future courts, opposing counsel, and bar disciplinary bodies. If “my sanction was only $1” is the story an attorney is telling themselves about a nominal-dollar order, they are misreading it.

How firms should price this on a portfolio basis

Two exposures, priced separately. First, the direct-fine exposure: bounded, largely predictable, $500 to $10,000 per incident in the modal case. Second, the adverse-costs exposure: unbounded above, sized by the opposing side’s reasonable fees on the sanctioned briefing, and correlated with case size and posture. A firm doing $500K+ commercial disputes should price this exposure at LiveVideo.AI or Couvrette magnitude, not at Mata magnitude. A firm doing consumer-defense or pro se-heavy work should price it at the low end of the distribution. Averaging the two exposures produces the wrong number for both practices.

Common questions

Are adverse-costs orders discretionary, or does Rule 11 require them?

Discretionary. The court decides whether the sanction includes a fee-shift component and whether to size it to the opposing party’s actual fees. In 2026 practice, courts are exercising that discretion toward fee-shifting more often when AI-fabricated citations are the underlying conduct.

Can we insure the adverse-costs exposure?

Ask your malpractice carrier — the answer depends on policy language. The 2026 renewal cycle is producing carrier-specific AI-drafting exclusions and disclosure requirements that materially change the coverage picture. Do not assume prior coverage extends without a specific written confirmation.

Verify a brief before you file it →

Deeper analysis

Operational note. The adverse-costs component is the number to build the internal risk model around. Direct fines are budget-line variance; adverse-costs orders are balance-sheet events. Every firm with contested civil practice should be running quarterly scenario reviews against a $50,000-$100,000 adverse-costs exposure per contested motion round, and staffing the pre-filing verification workflow accordingly. The 2026 orders make this the base-rate scenario, not the worst-case one.

The mechanical reason adverse-costs orders scale with case size is that they shift the reasonable attorney fees the opposing side incurred in responding to the fabricated citations back onto the sanctioned party. In a small pro se matter, that number is negligible. In a defended commercial case where opposing counsel has three attorneys billing at big-firm rates, the number is easily $50,000-$100,000 for a single motion round. That is why Couvrette (D. Or.) landed at $94,700 and LiveVideo.AI (S.D.N.Y.) at $80,056 while smaller-posture cases in the same 2026 batch cluster at $5,000-$15,000.

The malpractice-insurance implication is the underappreciated part. Direct sanctions of a few thousand dollars typically fall below the deductible on a professional liability policy and are absorbed by the attorney or firm. Adverse-costs orders in the $30,000-$100,000+ range breach the deductible and hit the carrier’s payout column. Once a claim is paid, the renewal premium repricing is measured in years. A single Couvrette-shape adverse-costs order can produce a decade of elevated malpractice premiums that dwarf the sanction itself.

The 2026 cluster is coherent enough that the industry can now underwrite it. Malpractice carriers are asking about firm AI-use policies and verifier workflow at renewal because the loss data on adverse-costs orders is legible. Firms that show a documented pre-filing verification workflow with contemporaneous logs are receiving modest premium credits in the 2026 renewal cycle; firms that cannot demonstrate one are seeing surcharges or exclusion language on AI-generated filing risk. The paper trail is now literally an asset.