Wadsworth v. Walmart: AI Hallucination Sanctions Hit Two Law Firms in Same Case | Advocate Prakhar

Case at a Glance

Background

A personal injury lawsuit against retail giant Walmart Inc. in the federal District of Wyoming became one of 2025’s most instructive AI accountability cases — not because of what Walmart did, but because of what the plaintiff’s legal teams did not do. Lawyers from two separate law firms, Morgan & Morgan and MX2.law, independently submitted briefs containing case citations that the court could not locate. When pressed, both firms admitted the citations had been generated by AI tools and had never been independently verified against actual legal databases.

The case is particularly notable because the AI verification failure occurred across two separate firms simultaneously — demonstrating that the problem of unverified AI citations is systemic, not isolated to technologically unsophisticated practitioners.

The AI Issue

The legal question before the court was whether attorneys who submitted AI-generated case citations — without independently confirming their existence in recognised legal databases — had breached their professional duties of competence and candor to the tribunal under Rule 11 of the Federal Rules of Civil Procedure and the court’s inherent supervisory authority.

A secondary issue was whether sanctions should extend to the law firms themselves, not just the individual attorneys who signed the briefs — raising important questions about vicarious responsibility for AI misuse in legal practice.

What the Court Decided

• Both Morgan & Morgan and MX2.law were sanctioned $5,000 each for submitting fabricated AI citations that did not correspond to real cases [violation of FRCP Rule 11 — duty to verify factual and legal accuracy].
• The court held that relying on AI-generated research without verification violated the fundamental duty of candor to the tribunal [candor toward the tribunal — Model Rule 3.3].
• Attorneys cannot delegate the verification obligation to an AI tool; the professional duty remains with the human lawyer at all times.
• The court rejected any suggestion that honest reliance on a new technology constitutes a mitigating circumstance — the verification duty predates AI and applies to all sources.
• Both firms were required to certify compliance with the court’s AI verification standards in all future filings.

“Counsel have an absolute duty to verify every legal citation before presenting it to this court. The use of artificial intelligence does not alter, diminish, or excuse that obligation.”

— US District Court, District of Wyoming, 2025

The India Angle

The Wadsworth v. Walmart sanctions resonate powerfully for Indian legal practice, where AI legal research tools are increasingly being adopted by junior advocates and law firms, often without established verification protocols.

Indian Law Equivalent

Under the Advocates Act, 1961 and the Bar Council of India Rules, an advocate who cites a non-existent case before a court may face contempt proceedings under the Contempt of Courts Act, 1971. Section 2(c) defines criminal contempt to include acts that scandalise or lower the authority of a court — fabricated citations, even if unintentional, risk crossing this threshold. Under the Code of Civil Procedure, 1908, Order XI Rule 1 and Order VI Rule 16 allow courts to strike out frivolous or scandalous material; a court discovering AI-hallucinated citations could invoke these provisions.

Bar Council Rules

Rule 14 of the BCI Standards of Professional Conduct and Etiquette mandates that an advocate shall not suppress material facts or knowingly make false statements to the court. Citing a non-existent case — regardless of whether AI generated it — potentially constitutes making a false statement. Rule 36 further requires advocates to maintain professional standards at all times.

Practical Advice for Indian Advocates

• Always verify every case citation on IndianKanoon, the Supreme Court’s official website, or the relevant High Court’s judgment portal before filing any pleading or written submission.
• Establish a two-person verification protocol in your office: the advocate who drafts using AI and a second person who independently checks every citation in an official database.
• Never copy AI output directly into a court document — treat AI suggestions as a first draft requiring human verification, not as a finished product.

Quick Takeaways

• Both law firms in the same case received AI hallucination sanctions simultaneously.
• “My AI tool generated it” is no defence against professional misconduct.
• Verify every citation in an official database before filing in any court.

Deep Dive: The Anatomy of a Dual-Firm AI Failure

Why Two Firms Made the Same Mistake

The Wadsworth case is structurally different from most AI hallucination sanctions because it involves not one but two independent law firms making the same category of error in the same litigation. This simultaneity is significant. It suggests that the adoption of AI research tools in litigation practice has outpaced the development of internal quality-control processes within law firms — a systemic failure, not an individual one.

Morgan & Morgan is a major plaintiffs’ firm with offices across the United States, handling thousands of personal injury cases. MX2.law is a smaller boutique firm. That firms of such different sizes and resources both failed to implement basic AI verification procedures illustrates that firm size and resources are not the determinative factor in AI compliance — training and culture are.

The Mechanics of AI Hallucination in Legal Research

AI large language models such as ChatGPT, Claude, and various legal-specific AI tools are trained on enormous datasets of text. They generate outputs by predicting what text is likely to follow the input prompt, based on statistical patterns in training data. When asked for case citations, an AI tool may produce plausible-sounding citations — with realistic party names, court abbreviations, and year references — that simply do not correspond to any actual decided case.

The AI is not lying in any meaningful sense; it is generating statistically probable outputs. But legal citation is a domain where plausibility is not the same as accuracy. A hallucinated citation to Smith v. Jones, 892 F.3d 144 (8th Cir. 2018) may look entirely legitimate but be wholly invented.

FRCP Rule 11 and the Duty of Verification

Federal Rule of Civil Procedure 11(b) requires that when an attorney signs and files a document with the court, the attorney certifies — by that very act of signing — that to the best of their knowledge, formed after reasonable inquiry, the legal contentions in the document are warranted by existing law or by a nonfrivolous argument for extending existing law. The phrase “reasonable inquiry” is crucial: it is not satisfied by reliance on AI output alone. The Rule 11 standard predates AI by decades but applies with full force to AI-generated content.

Courts have consistently held that Rule 11 imposes an affirmative duty to investigate, not merely an obligation to avoid deliberate falsehood. Reasonable inquiry into a case citation means actually looking the case up in a verified legal database — Westlaw, LexisNexis, CourtListener, or equivalent.

The Trajectory of AI Hallucination Sanctions in the US

Wadsworth v. Walmart falls in a line of increasingly serious AI hallucination sanctions that began with Mata v. Avianca (SDNY, 2023), where Judge Castel imposed $5,000 sanctions on attorneys who cited ChatGPT-hallucinated cases. Since Mata, courts across the United States have imposed sanctions ranging from $1,000 to $110,000 (Oregon, 2025), and at least one attorney has been referred for bar discipline as a consequence of AI-generated false citations.

The Wadsworth case is notable because it demonstrates the post-Mata period: the profession has had two years to absorb the lesson of Mata and other early cases, and attorneys are still making the same verification errors. Courts are responding with less patience and equivalent or higher sanction amounts.

Global Comparison

Courts in Canada, the United Kingdom, and Australia have begun issuing practice directions and guidance notes on AI use in litigation. Canadian courts have been particularly active, with several provincial superior courts publishing specific rules about AI disclosure and verification requirements. The Wadsworth sanctions fit within a global trend of courts asserting authority over AI-assisted legal practice through the existing mechanisms of professional conduct rules and inherent supervisory powers.

In India, no equivalent reported case of AI hallucination sanctions exists as of mid-2026 — but the foundational professional duties that would support such sanctions are already present in the existing framework of the Advocates Act, Contempt of Courts Act, and Bar Council Rules. Indian courts have not yet been confronted with the question but may well be, as AI legal research tools become more accessible to practitioners across the country.

What to Watch For Next

The critical development to watch in AI hallucination jurisprudence is the emergence of bar discipline proceedings arising directly from AI sanction orders. Several US courts have referred attorneys to state bar authorities after imposing monetary sanctions. The next major development will be disbarment or suspension proceedings based on AI misuse — at which point the professional consequences will become existential rather than merely financial.