October 18, 2026 (in 16 days): Tenth Circuit: comments close on 2027 rules, incl. proposed Rule 46.5(B)(4) generative-AI review →

Daghra v. Hinkley

U.S. District Court, Western District of Michigan, Southern Division · W.D. Mich. · Michigan bar guidance

Conduct

In a habeas response, the Government quoted and cited a Sixth Circuit case, Taylor v. Hott, that the court could not locate; the cited page belongs to an unrelated opinion.

Consequence

No sanction "presently"; the court said the Government "must ensure" its future filings contain no non-existent case law. The petition was dismissed as moot.

Lesson

Here the court found a different, unrelated opinion at a government filing's cited page, and addressed the citation even though the petition had become moot.

Warning

Verified September 30, 2026

Citation
Daghra v. Hinkley, No. 1:26-cv-1429, Order (W.D. Mich. July 16, 2026) (Jarbou, C.J.), ECF No. 11
Decided
July 16, 2026

Summary

Izzeddin Ahmed Abdulghaffar Daghra, an Immigration and Customs Enforcement detainee, petitioned for a writ of habeas corpus under 28 U.S.C. § 2241, arguing that the 90-day automatic stay of his $35,000 bond order under 8 C.F.R. § 1003.19(i) violated due process. In its response to the court's initial order to show cause (ECF No. 5), the Government wrote that the Sixth Circuit had "reiterated that § 1226(e) bars challenges that 'ask the court to reweigh the evidence underlying a bond decision or second-guess the Immigration Judge's discretionary judgment'" and cited "Taylor v. Hott, 724 F. App'x 387, 392 (6th Cir. 2018)." The court found that Taylor v. Hott "is not located at the identified page of the Federal Appendix"; page 387 falls within a different opinion, Atkins v. CGI Techs. & Sols., Inc., 724 F. App'x 383 (6th Cir. 2018), which concerns commercial arbitration. The court "was unable to identify a Sixth Circuit case with the caption Taylor v. Hott, or any federal case containing the quoted language," and concluded that "it seems this citation was likely produced by generative artificial intelligence." The order does not name the Government's counsel.

AI tool:
Unidentified (the court wrote that "it seems this citation was likely produced by generative artificial intelligence"; the order names no tool and records no explanation from the Government)
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 did the court warn about?

Chief Judge Hala Y. Jarbou dismissed the petition without prejudice as moot, because the automatic stay had expired and the Government represented that the bond order was back in effect. On the citation, the court stated that "any attorney who uses AI must scrupulously review its work product to ensure that the cited cases exist and that the citations accurately and fairly represent the underlying case law," and that "[t]he duty of candor towards this tribunal demands no less." It then wrote: "Although the Court will not presently impose sanctions for this conduct, it goes without saying that the Government must ensure its future filings with this Court do not include non-existent case law." The order imposes no monetary or other sanction, orders no further response on the citation, and makes no referral. A separate judgment was entered the same day.

Why does Daghra v. Hinkley matter for law firms using AI?

Daghra shows a court raising a suspected AI-generated citation in an order that otherwise disposed of the case on mootness. The habeas petition challenged the automatic stay of an immigration judge’s bond order; by the time the court ruled, the stay had expired and the Government represented that the petitioner would be released on posting bond. Chief Judge Jarbou nonetheless wrote that there was “one additional issue in this case that the Court must address”: the Government’s response had quoted a Sixth Circuit decision, Taylor v. Hott, that the court could not find at the cited Federal Appendix page, or as any Sixth Circuit case by that name.

The court’s reasoning rests on what its own research showed. The cited page, 724 F. App’x 387, sits inside an unrelated commercial-arbitration opinion, and no federal case the court could identify contains the quoted language. From that, the court inferred that “it seems this citation was likely produced by generative artificial intelligence.” The order does not report any explanation from the Government, and it does not find that AI was in fact used. It quotes another Western District of Michigan decision, Lothamer Tax Resolution, Inc. v. Kimmel (itself quoting Evans v. Robertson, E.D. Mich.), for the observation that “courts have seen a rash of cases in which both counsel and pro se litigants have cited such fake, hallucinated cases in their briefs.”

The consequence was a warning, not a sanction. The court said it “will not presently impose sanctions for this conduct,” and that the Government “must ensure its future filings with this Court do not include non-existent case law.” Firms documenting compliance may wish to consider that the court tied its expectation to the duty of candor and to checking that cited cases exist and are accurately and fairly represented, and that it treated the problem as worth addressing even though the petition itself was moot.

Implications for your firm

Operational steps a firm reading this case may wish to consider documenting. Strategic and rule-application calls belong to your firm's attorneys.

  • Verify that every quotation attributed to a case appears in that case at the cited page. The court checked the reporter page, found a different opinion there, and searched for the quoted language in any federal case.
  • Consider that a citation problem can be addressed even after a case becomes moot. The court addressed the citation in the same order that dismissed the petition as moot.
  • Consider that the court grounded its warning in the duty of candor: it said any attorney who uses AI must review its work product so that cited cases exist and are fairly represented.

Sources

Primary sources