How Not to Get Sanctioned
New York Law Journal, Litigation
Immediately post-pandemic, the advice that attorneys should double-check legal research conducted by a large language model would have had many of us asking what, exactly, is a large language model. By June 2023, the principle that attorneys have a “gatekeeping role” to “ensure the accuracy of their filings” against “non-existent judicial opinions with fake quotes and citations” created by AI was accepted wisdom. Mata v. Avianca, Inc., 678 F.Supp.3d 443, 448 (S.D.N.Y. 2023).
It is the rare New York Law Journal reader who would blindly submit a brief “peppered with false citations” created by “generative artificial intelligence.” Flycatcher Corp. v. Affable Ave. LLC, No. 24 Civ. 9429 (KPF), 2026 U.S. Dist. LEXIS 23980 at *1 (S.D.N.Y. Feb. 6, 2026). To say in 2026 that doing so invites sanctions is banal.
