A Michigan federal judge has issued a warning to the U.S. Justice Department after discovering that a government court filing in an immigration case had cited a legal case that doesn’t exist.

Judge Hala Y Jarbou for the Western District of Michigan passed an order in an immigration case of petitioner Izzeddin Ahmed Abdul Gaffar Daghra on Thursday, wherein a discrepancy related to a cited case in the filing was also discussed by the judge.

The government filing cited a Sixth Circuit case, Taylor v. Hott, as support for the government’s position on immigration bond decisions.

But Judge Jarbou said the cited case could not be found anywhere in federal courts history and the page number cited in the citation was for a different case, which dealt with commercial arbitration and not immigration law.

The judge also indicated that the court could not find any ruling that matched the name given in this case or had any of the legal phrasing used.

Jarbou mentioned that the reference was “most probably created by generative artificial intelligence,” but the court could not reach any final conclusion on this matter. The judge also mentioned that different courts across the US began noticing issues concerning fake cases produced by AI.

The judge decided not to impose penalties as of now, but made it clear that lawyers who use AI need to double-check their information before submitting any documents to the court.

“It should be evident that any attorneys who utilize AI need to carefully verify the information used,” Jarbou said, making it clear that lawyers are obliged to be truthful and accurate with regard to the court.

The case concerned an immigrant who was contesting the delay concerning bond approval. After determining that the case was moot because of government’s statement that the detainee could post bond, the court dismissed the case.

Also Read: When Judge Meets the Machine: Can AI Help Deliver Justice Without Losing Humanity?