Courts on AI / Fabricated or misquoted citations

Hack v. Preston

United States District Court for the District of Utah · Federal district courts · UT · 2026-08-07 · No. 4:25-cv-00096-DN · order

Fabricated or misquoted citations Self-represented litigants using AI Court rules adopted by opinion or orderChatGPT

The District of Utah addresses Jayne Stephens’s declaration that ChatGPT assisted her filings. The order notes a mischaracterized assertion based on an improperly cited Tenth Circuit case, requires future assistance disclosures, and cautions that GenAI-related Rule 11 violations may be sanctioned.

Disposition: The court required future assistance declarations and cautioned the self-represented defendant about Rule 11 sanctions.
Recorded conduct: Self-represented defendant used ChatGPT and filed a reply with an improperly cited and mischaracterized Tenth Circuit case. — outcome: warning; actor: litigant in person
Canonical record: Incident tracker entry — verified outcome, regulator disposition and related rules

What the court wrote

Verbatim from the court's document (public domain). Ellipses mark omitted text; nothing is paraphrased.

Ms. Stephens filed the Declaration on August 3, 2026, and declared, under penalty of perjury, that she had used a generative artificial intelligence (“GenAI”) tool known as ChatGPT. She confirms that she utilized ChatGPT “for assistance with research, organization, drafting, editing, and formatting.” Her candor is appreciated, but necessitates a caution. ... Ms. Stephens has acknowledged using ChatGPT, a GenAI tool that is widely known for its potential to produce inaccurate or fabricated information, commonly referred to as “AI hallucinations.” Although the use of GenAI is not inherently improper, it presents a risk that filings may contain unsupported or inaccurate legal authority if not carefully verified. In her reply memorandum, supporting her motion to dismiss for insufficient service, Ms. Stephens includes a mischaracterized legal assertion relying on a case that was improperly cited. Briefing has not yet crossed the threshold to warrant sanctions, but the risks of AI hallucinations in future filings concern the court. Accordingly, this order serves as a caution that sanctions under Rule 11 of the Federal Rules of Civil Procedure may be imposed if the use of GenAI results in violations of the Federal Rules of Civil Procedure.

Authorities cited in the passage: Picon-Diaz v. Bondi, No. 25-9530, 2026 WL 412348 (10th Cir. Feb. 13, 2026) · Moore v. City of Del City, No. 25-6002, 2025 WL 3471341 (10th Cir. Dec. 3, 2025) · Hukill v. Oklahoma Native Am. Domestic Violence Coal. · Fed. R. Civ. P. 11

Court document: damiencharlotin.com · archived copy
Read: 2026-09-07 · court document read from a public mirror; official copy pendingprovisional
Found via: charlotin-cc0
Notes: Official source not obtained after CourtListener throttling; read the court-authored document from the Charlotin mirror and marked link-only.

United States court documents are public domain; the topic coding, disposition line and summary are SafeLegalAI's reading of the document for comparison, not a finding about any person named in it. This record summarises a public court document or official publication and reports what it says; SafeLegalAI makes no findings of its own. Individuals and organisations are named only as they appear in that public record. Anyone named may send a right of reply, correction or contextual update (for example an appeal, vacatur or later disposition) via /report; we add a dated note where appropriate. Reply or correct this record.

compiled by the Cognesio LLP editorial teammethodas of 2026-09-07v0.1.1coding CC BY 4.0JSON Hugging Face GitHub