Litigation News

US court sanctions lawyers for filing AI-generated fake case laws in patent suit

The court said that while the use of generative AI in the legal profession is not inherently problematic, the failure to verify its output constitutes a serious breach of professional standards.

S N Thyagarajan

A United States district court in Kansas has imposed sanctions on lawyers representing Lexos Media IP, LLC after finding that court filings contained fabricated case citations, non-existent quotations and misstatements of law generated using artificial intelligence (AI) tools without verification.

In a detailed memorandum and order dated February 2, 2026, Judge Julie Robinson held that all five attorneys who signed the filings violated Rule 11 of the Federal Rules of Civil Procedure, which requires lawyers to ensure that legal contentions are warranted by existing law after a reasonable inquiry.

The sanctions arose during a patent infringement lawsuit filed by Lexos Media IP against Overstock.com. On July 7, 2025, Lexos' legal team filed two response briefs: one opposing a motion to exclude their technical expert and another responding to a summary judgment motion. Overstock’s subsequent reply brief alerted the court that several citations in these filings were "defective".

The court determined that the attorneys violated Federal Rule of Civil Procedure 11(b) by failing to conduct a "reasonable inquiry" into the accuracy of the law cited in their briefs. The offending documents, filed on July 7, 2025, included:

Non-existent Cases: One filing twice cited Hockett v. City of Topeka (2020), a case that the court found does not exist in any federal database.

Fabricated quotations: The briefs included specific "quotes" attributed to actual Federal Circuit cases, such as Liquid Dynamics Corp. v. Vaughan Co. and Apple Inc. v. Motorola, Inc, that were not found in those decisions.

Misrepresented authority: Several cases were cited for legal propositions they did not support. For instance, Flexuspine, Inc. v. Globus Medical, Inc. was cited regarding expert admissibility, but the court found it contained no such discussion on the cited page.

Attorney Sandeep Seth admitted to being the primary drafter and using ChatGPT as a "shortcut" to find case law while dealing with personal family medical emergencies. Seth claimed he was "untrained" on how to use generative AI and was a "novice" who was unaware of the potential for the software to "hallucinate" or invent information. Although Seth took primary responsibility, the court ruled that all signing attorneys had a duty to verify the contents of their filings. It held,

"While the Court is sympathetic to Mr. Seth’s family circumstances, it cannot ignore that there were obvious options open to him short of using a publicly-available generative AI search tool that he had never used before to research his brief without verifying the accuracy of that research."

The lawyers were handed fines ranging from $5,000 to $1,000, as well as other disciplinary action, based on their roles and the severity of the oversight.

Judge Robinson emphasised that while the use of generative AI in the legal profession is not inherently problematic, the failure to verify its output constitutes a serious breach of professional standards. The court noted that the errors caused significant prejudice to the defendant and wasted judicial resources.

[Read Judgment]

Lexos Vs Overstock.pdf
Preview

Karnataka HC moots guidelines to curb litigants from filing bogus cases against opposite side's lawyers

Allahabad High Court orders Consortium of NLUs to revise CLAT-2026 merit list

Supreme Court quashes Madras High Court's late evening stay on laws that got Governor's deemed assent

Foreign Exchange Management (Guarantees) Regulations, 2026: A brief overview

Sonam Wangchuk not in good condition: Supreme Court urges Centre to review detention

SCROLL FOR NEXT