Case D-53 · AI Sanctions Case Tracker

Akerlund v. Atlas Air, Inc.

U.S. Court of Appeals for the Eleventh Circuit, 181 F.4th 1200 (No. 24-11033, opinion by Judge Britt C. Grant, filed July 10, 2026)

The case

Akerlund v. Atlas Air, Inc., 181 F.4th 1200 (11th Cir. 2026) (No. 24-11033, Judge Britt C. Grant). Primary opinion: https://govinfo.gov/content/pkg/USCOURTS-ca11-24-11033/pdf/USCOURTS-ca11-24-11033-0.pdf ; reported: https://www.reuters.com/legal/litigation/us-appeals-court-rebukes-lawyer-over-fake-hallucinated-case-citations-2026-07-10/ ; https://news.bloombergtax.com/health-law-and-business/ex-florida-congress-candidate-slammed-by-court-over-ai-in-briefs

What happened

Anthony F. Sabatini (Sabatini Law Firm, extra-document detail, per reporting; the opinion names no firm) filed multiple briefs containing fake AI-hallucinated citations: the opening brief alone had at least eight. When confronted, he withdrew eight more fabricated citations.

The rule and the sanction

The panel issued a sharp public rebuke ("no substitute for actual intelligence"), and the chief judge was to refer Sabatini to the Committee on Lawyer Qualifications and Conduct. Like Parnell (D-52), decided the same week, this shows the Eleventh Circuit's 2026 pattern of routing AI-citation misconduct into its formal conduct pipeline. Sabatini had apologized the prior year for "erroneous or unverifiable" citations (Reuters).

How it could have been avoided

Withdrawing fabricated citations after being caught does not cure the misconduct: the confrontation becomes part of the record and compounds it. Fakes spread across multiple briefs read as a practice, not a one-off mistake. And profile cuts against you: a publicly known lawyer gets press coverage, not the benefit of the doubt. A prior apology without changed practices aggravates the outcome; after any AI-related incident, document and implement a new verification protocol.

Sources