Sanction teardown · E.D. New York, USA · 2026-03-31
Evgeny Yaroshevsky v. Karlin et al.
What happened
In E.D. New York, USA, a filing relied on Gemini to help draft legal argument. The court identified the following problems with the citations in that filing:
-
Misrepresented (Other)Plaintiff admitted using Google Gemini to obtain citations and failed to verify them before filing, contributing to the submission of fabricated and inaccurate authorities.
-
Fabricated (Case Law)Multiple authorities cited by plaintiff did not exist; defendants identified at least five nonexistent cases in the complaint and five in other filings and the Court confirmed some citations were fabricated.
-
False Quotes (Case Law)Numerous quotations attributed to cases could not be found in the opinions (Defendants identified ~13 instances; the Court found at least two such misquotes earlier).
Which AI tool
Gemini. Note: Charlotin's public database records tool attribution only where a court order, brief, or reporting on the matter states it explicitly; "unidentified" or "implied" means the record indicates AI use but does not name a specific product — we do not guess.
Outcome
Recommended adverse costs order
Additional detail
R&R was adopted later on (see here).
How Citation Safe would have caught this
Citation Safe runs three deterministic layers before a brief is filed: (1) does the citation exist against CourtListener's database of published opinions, (2) if quoted, does that exact language appear in the source, (3) does the cited case actually support the proposition it is cited for. Fabricated case citations fail Layer 1. Fabricated or misattributed quotations fail Layer 2 even when the underlying case is real. Misrepresented holdings — a real case cited for a proposition it does not support — are the target of Layer 3. None of these checks involve asking another language model whether the citation looks right; they are lookups and text-matches against the actual source, which is why a hallucinated citation has to survive a direct lookup against the authoritative source — not another model's opinion — to earn a VERIFIED stamp; our measured false-verify rate is published live at /quality.
Check a brief before you file it → · See our live false-verify rate
Source: https://www.damiencharlotin.com/documents/1592/Yaroshevsky_v._Karlin_USA_23_February_2026.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).