Wadsworth v. Walmart: AI hallucinated eight of nine cases in a Morgan and Morgan brief, and a federal judge fined the lawyers $5,000 and revoked one pro hac vice

The problem

Generative AI marketed for legal research can fabricate authority that looks real. In Wadsworth v. Walmart Inc. (D. Wyo., No. 2:23-CV-118-KHR), a product-liability suit, plaintiffs’ attorneys from the national firm Morgan and Morgan filed motions in limine on 22 January 2025 that, in the court’s words, “cited nine cases, but eight did not exist” (source). At a hearing the attorneys admitted the missing cases had been hallucinated by an AI tool (source).

What was built

The tool was Morgan and Morgan’s own in-house generative-AI legal-research platform. Per the order, drafter Rudwin Ayala “uploaded the brief onto ‘MX2.law’ to add case law,” which the court describes as “an in-house database launched by Mr. Ayala and Mr. Morgan’s firm, Morgan & Morgan” (source). Ayala included the fabricated citations without verifying them, and the order records that “this was his first time ever using AI in such a way” (source). LawSites reported the same platform: “Ayala used his firm’s in-house AI platform, MX2.law, to generate case law when drafting the motions” (source).

The outcome

On 24 February 2025, U.S. District Judge Kelly H. Rankin found all three attorneys violated Rule 11(b) and imposed sanctions (source).

The drafter: $3,000 and a revoked pro hac vice. Per LawSites, “Rudwin Ayala had his pro hac vice admission revoked and was ordered to pay a $3,000 fine” (source), a disposition Reason’s Volokh Conspiracy took from the same order (source).

The two other lawyers: $1,000 each. Supervising attorney T. Michael Morgan and local counsel Taly Goody “were each fined $1,000” (source). Each attorney “shall pay the amount into the Registry of this Court within fourteen days of this Order” (source).

The total: $5,000 across three lawyers. The order lists the three penalties individually; the sum is not stated as a separate figure in the order, and is presented here as the arithmetic total of the three, not a fourth court-stated number (source).

The fabrication: 8 of 9 cited cases did not exist. The motions in limine “cited nine cases, but eight did not exist” (source), a count LawSites relayed as “motions in limine that cited eight non-existent cases” (source).

The firms were not sanctioned. Judge Rankin “declined to sanction the attorneys’ law firm, noting that Morgan & Morgan had trained its employees not to use AI software in the way Ayala had and had since implemented an additional acknowledgment requiring users to independently verify AI-generated information” (source).

How this was verified

This case carries a green badge as a court-adjudicated primary record, not a deployment win. The figures are stated by an independent federal court in a signed, publicly filed order (ECF No. 181, signed 24 February 2025), whose full text was re-fetched and re-checked figure by figure on 2026-08-15, and two newsrooms independent of each other, of the court and of the firms reported the same figures the next day: LawSites (Bob Ambrogi) and Reason’s Volokh Conspiracy (Eugene Volokh) (source). Because the court, not the sanctioned parties, states the numbers, no client-confirmation call is needed or used.

The honest limit: the order itself is fully verifiable, but the post-order docket facts (the Registry payment of the three penalties, and whether any appeal was taken) are PACER and CourtListener-authentication-gated and were not re-verifiable in this pass, so none of those specifics is asserted here. Bloomberg Law reported the same $3,000 and $1,000 figures but sits behind a paywall, so it is treated as headline-only corroboration rather than a byte-checkable secondary.

Related case files

The same failure, an attorney filing AI-fabricated citations and drawing a Rule 11 sanction, runs through Gauthier v. Goodyear, where a single Texas lawyer drew a $2,000 fine and a generative-AI CLE order, a useful contrast with Ayala, who did not correct the brief until the court acted. Mid-Central v. HoosierVac is the closest sibling on the tool, an in-house or self-served AI feeding fabricated citations across multiple briefs for a $6,000 Rule 11 sanction. Coomer v. Lindell shows the same per-attorney apportionment of blame, with the MyPillow lawyers each sanctioned $3,000 for a brief riddled with AI-defective citations.

Sources

Cited in this case file. Tier 1 = the court’s own signed order; Tier 2 = independent press naming the parties; Tier 3 = paywalled or aggregated report. Each figure was checked against the live source on 2026-08-15.

1. U.S. District Court, D. Wyoming, “Order on Sanctions,” Wadsworth v. Walmart Inc., No. 2:23-CV-118-KHR, ECF No. 181, signed by Judge Kelly H. Rankin, decided 2025-02-24 (Tier 1, primary court order; full text hosted at FindLaw). https://caselaw.findlaw.com/court/us-dis-crt-d-wyo/117003959.html
2. LawSites, Bob Ambrogi, “Federal Judge Sanctions Morgan & Morgan Attorneys for AI-Generated Fake Cases in Court Filing,” 2025-02-25 (Tier 2, independent legal press; confirms the $3,000 and $1,000 fines, the pro hac vice revocation, MX2.law, and the firm’s non-sanction). https://www.lawnext.com/2025/02/federal-judge-sanctions-morgan-morgan-attorneys-for-ai-generated-fake-cases-in-court-filing.html
3. Reason, The Volokh Conspiracy, Eugene Volokh, “Sanctions on Lawyers for Filing Motion Containing AI-Hallucinated Cases,” 2025-02-25 (Tier 2, independent; quotes the order on MX2.law and the penalties). https://reason.com/volokh/2025/02/25/sanctions-on-lawyers-for-filing-motion-containing-ai-hallucinated-cases/
4. Bloomberg Law, Bernie Pazanowski, “Morgan & Morgan Lawyers Fined for ‘Hallucinated’ AI Citations,” 2025-02-25 (Tier 3, independent but paywalled; headline-only corroboration of the same figures). https://news.bloomberglaw.com/litigation/morgan-morgan-lawyers-fined-for-hallucinated-ai-citations