June 1, 2026 (in 3 days): New York: 22 NYCRR Part 161 takes effect, system-wide AI policy for all UCS courts

Dec v. Mullin

U.S. Court of Appeals for the Seventh Circuit · 7th Cir. · Illinois bar guidance , Indiana bar guidance , Wisconsin bar guidance

Court sanction

Verified April 26, 2026

Citation
Dec v. Mullin, No. 25-2417 (7th Cir. Mar. 30, 2026) (Brennan, J.)
Decided
March 30, 2026

Summary

Petition for review of a Department of Homeland Security denial of an immigration waiver. The Seventh Circuit held it lacked jurisdiction to review the discretionary waiver decision and dismissed the petition. Separately, the panel addressed petitioner's opening brief, which cited two non-existent cases and recounted a fabricated quotation. The court characterized these as "tell-tale signs" of AI hallucinations, where a large language model generates nonexistent citations. The opinion notes that the court's "concern lies with trained lawyers failing to check the accuracy of legal citations and quotations in their filings," and treats the duty as technology-agnostic.

AI tool:
AI-generated (court found error not knowing/intentional)
Sanction amount:
None (public admonishment, no monetary sanction)
This case summary is informational only. Verify the underlying opinion or order against the primary source before relying on it in any filing or client matter.

What sanction did the court impose?

Petition for review dismissed for lack of jurisdiction. Petitioner's counsel publicly admonished. The panel declined to impose formal sanctions, citing counsel's roughly ten years of practice, her heavy caseload tied to the Chicago immigration surge, her acceptance of responsibility, and her sincere apology, and finding that the false citations were not made knowingly or intentionally.

Why does Dec v. Mullin matter for law firms using AI?

Dec is the first federal circuit-level admonishment in this product’s tracker for AI-style hallucinated citations in an appellate brief, and the Seventh Circuit’s framing matters for managing partners. The court did not anchor its concern to ChatGPT, Westlaw, or any specific tool. It anchored the duty to verify on the lawyer, regardless of the research method. Counsel’s volume of immigration work and apology bought leniency on sanctions but did not avoid public admonishment in a published opinion. For firms documenting AI use in appellate practice, the takeaway is that “we were busy” is not a defense at the Court of Appeals level, and that citation verification against primary sources is treated as a baseline professional obligation rather than a tool-specific best practice.

Sources

Primary sources

Further reading

Source PDF is a Westlaw printout mirrored from the Damien Charlotin hallucination database. We are working to add the underlying court docket (PACER, CourtListener, or court website) as a second source.

Unverified claims:
  • Petitioner's counsel's name was not extracted from the opinion text during verification (PDF stream not directly readable via WebFetch); secondary coverage describes her as a Chicago immigration practitioner with about ten years of experience but does not name her.
  • The two non-existent case names and the exact text of the fabricated quotation were not extracted from the opinion during verification.
  • Full panel composition beyond authoring Judge Michael B. Brennan was not confirmed.
  • The case caption uses "Mullin" (the named DHS Secretary in his official capacity) as respondent; some coverage refers to the matter as "Dec v. DHS."