A federal judge in Michigan has expressed concern over a filing from the Trump administration’s Department of Justice, accusing it of using an AI-generated citation for a nonexistent court decision. This emerged from legal proceedings in an immigration detention case.
Judge’s Response to AI Citations
Chief U.S. District Judge Hala Y. Jarbou, presiding over the Western District of Michigan, issued a three-page order on July 16. The order pertained to a habeas corpus case involving an Immigration and Customs Enforcement (ICE) detainee challenging the government’s automatic stay of a bond order. While the case was dismissed as moot since the stay expired and the detainee was cleared for release upon posting a $35,000 bond, significant attention was drawn to the questionable citation presented by government lawyers.
According to the judge’s order, the Department of Justice claimed that the Sixth Circuit ruled federal courts could not reassess evidence or question an immigration judge’s discretionary judgment, referencing a nonexistent case, Taylor v. Hott.
Judiciary Tensions and Scrutiny
This issue is not isolated, occurring amid growing friction between the judiciary and the Trump administration. Federal judges have been increasingly critical of actions taken by the administration, questioning their validity and accuracy.
Judge Jarbou underscored the responsibility that attorneys have when using AI tools, stressing the need for meticulous verification of work before submissions. She warned that AI-driven legal research has led to a surge in cases featuring fabricated authorities, with courts encountering briefs laden with “fake, hallucinated cases” previously unseen.
While refraining from imposing sanctions, the judge emphasized the necessity for future filings to be free of false citations and nonexistent legal references.
Past Incidents of AI Legal Filings
Michigan is not the first state to scrutinize AI-generated legal filings. In April, U.S. Magistrate Judge Robert T. Numbers II of North Carolina admonished former federal prosecutor Rudy Renfer. Renfer had admitted using generative AI for a court brief that included fabricated quotations, incorrect citations, and inaccuracies in case law representation.

Luigi Mangione Pleads Guilty in UnitedHealthcare CEO Brian Thompson’s Killing
Kim Mathers Jailed for Violating Drug and Alcohol Tests
Judge Blocks Part of Trump’s Mail-In Voting Executive Order
Lindsay Clancy’s Trial Case and Developments
ICE Agent Pleads Guilty to Battery Charge
DHS Demands California Hand Over Illegal Immigrant Accused of Killing Retiree