Incident Tracker / United States (federal)
Hill v. Foundation Media LLC, No. 1:25-cv-05947 (AS) (S.D.N.Y. 2026)
·United States District Court for the Southern District of New York·US-FEDWarning
In New York federal court, Judge Arun Subramanian ordered defence counsel to explain AI-affected quotations in a response brief, took no further action, and called the episode a wakeup call. Counsel's 1 October letter disclosed that Harvey, including updates referred to as Harvey 2.0, was used to prepare the response.
- Court
- United States District Court for the Southern District of New York
- Date
- 2026-09-29
- Who used the AI
- Lawyer
- AI tool
- Harvey (original version, potentially supplemented by updates collectively referred to as Harvey 2.0)
- Conduct
- The defendant's response to the pro se plaintiff's objection to a report and recommendation attributed quotations to Piazza and included a citation to Zhao that prompted the court to order lead counsel to declare whether unchecked AI produced them. Counsel then disclosed by letter that Harvey was used to prepare the response.
- Outcome
- Warning
- Penalty
- No monetary penalty recorded
- Regulatory outcome
- No separate regulatory disposition recorded
- Status
- Verified against listed sources
- Last checked
- 2026-10-05
What was fabricated or misused
- Quotations attributed to Piazza in Foundation's response to the plaintiff's objection
- A citation to Zhao in the same response
What the court said
“lead counsel for Foundation is ordered to file a declaration under 28 U.S.C. § 1746, on or before September 28, 2026, stating (1) whether or not AI was used in the preparation of Foundation’s response to Hill’s objection and (2) whether the quotations attributed to Piazza and the citation to Zhao were a result of Foundation’s unchecked use of AI.”
“No further action will be taken in this instance, and the Court hopes that no punitive action will be taken as to the lawyer who principally worked on Foundation’s response.”
“everyone should be aware that even the most advanced LLMs are prone to hallucination, and sometimes it will be hard to catch.”
“the original version of Harvey, as potentially supplemented by one or more updates referred to collectively as Harvey 2.0, was used in preparing Defendant’s response to Plaintiff’s objection.”
Timeline
- Magistrate judge recommended dismissal of the complaint with prejudice.
- District judge adopted the recommendation and ordered lead counsel to declare whether AI produced the Piazza quotations and Zhao citation.
- Court took no further action, set out four lessons, and ordered disclosure of the brand and version of AI used.
- Counsel's letter disclosed that Harvey was used.
Why this case matters
The court made the defence name the commercial product behind suspect quotations, and it was a legal-specific platform rather than a general chatbot; as of 5 October 2026 this is the only tracker record naming Harvey as the tool used. The order also addresses clients who require AI drafting to cut cost and warns that over-reliance may stunt junior lawyers' training. The record does not establish that Harvey caused the errors; it records counsel's disclosure.
Practice note
Lead counsel should confirm that every citation, quotation and legal proposition in an AI-assisted brief is checked before filing, as the order directs. Firms using a legal AI platform should be ready to tell a court which product and version produced a draft. The court suggested the first draft of a brief be written without AI where junior lawyers are being trained.
Primary sources for this incident
- Hill v. Foundation Media LLC, S.D.N.Y. No. 25-cv-5947, Order (Dkt. 57), 29 Sep 2026 — CourtListener RECAP copy
- Defendant's counsel letter disclosing the AI tool used (Dkt. 58), 1 Oct 2026 — CourtListener RECAP copy
- Order Adopting Report and Recommendation (Dkt. 54), 22 Sep 2026, ordering counsel's AI declaration — public copy (Charlotin document mirror)
Related regulation records for this incident
- Individual Practices in Civil Cases: Use of ChatGPT and Other ToolsSouthern District of New York (Judge Subramanian) · US-FED
Similar incidents on the record
- G.P. v. Jose Hernandez Arevalo, No. C26-284-KKE (W.D. Wash.)2026-09-08 · US District Court, Western District of Washington · WarningIn the US District Court for the Western District of Washington, plaintiff's counsel cited a nonexistent federal case in briefing against Turo. Judge Evanson described the citation as likely AI-fabricated, compelled arb…
- Williams v. Dallas-Fort Worth International Airport Board, No. 4:26-cv-00433-P (N.D. Tex.)2026-08-27 · US District Court, Northern District of Texas · WarningIn the Northern District of Texas, plaintiffs' counsel cited nonexistent and materially inapposite cases while trying to defeat qualified immunity. Judge Pittman called the effort borderline sanctionable, warned that ge…
- Shelbert v. Baxter International, Inc., No. 1:26-cv-06266 (N.D. Ill.)2026-08-26 · US District Court, Northern District of Illinois · WarningIn the Northern District of Illinois, Tyler Shelbert's lawyer filed a motion to strike that relied on a false reading of Instituto Nacional de Comercializacion Agricola. Chief Judge Kendall inferred AI-assisted drafting…
- Disruptive Resources, LLC v. Ballistic Barrier Products Inc., No. 1:24-cv-00321-JCG (D. Del.)2026-08-20 · US District Court for the District of Delaware (Judge Jennifer Choe-Groves, sitting by designation) · WarningIn a patent case in the US District Court for the District of Delaware, plaintiff's counsel promptly disclosed that a Strongsuit-assisted claim-construction brief contained errors. Judge Jennifer Choe-Groves discharged…
Cite this record
SafeLegalAI Global Legal AI Incident Tracker, “Hill v. Foundation Media LLC, No. 1:25-cv-05947 (AS) (S.D.N.Y. 2026)”, safelegalai.com/tracker/hill-v-foundation-media (accessed 2026-10-05). Data: CC BY 4.0.