‘Avalanche’ of likely AI-generated filings decried by MD judge in pro se case

Listen to this article

Suspecting that a recent deluge of motions might have been drafted by artificial intelligence, a federal judge in Maryland threatened sanctions against two self-represented litigants from Germany if they filed more “frivolous” filings or fabricated legal citations.

“Although the use of AI tools for legal research or drafting is not inherently improper, Plaintiffs’ avalanche of filings here, whether AI-generated or not, have burdened the Court, advanced frivolous arguments, misstated the record, and fabricated legal authorities,” U.S. District Court for Maryland Judge Adam B. Abelson wrote in his Monday opinion.

He noted that the plaintiffs, who are seeking to toss a previous settlement agreement, had “filed an unusually high number of motions in a short amount of time,” including a motion to exceed the maximum word and page count for filings, an emergency motion to stay and a motion for a “Complete Reset of these Proceedings” that compared the court to the notorious “Star Chamber” that conducted secret proceedings and punished dissent against the English monarchy.

The ruling came as judges grapple with the role of generative AI, which can rely on “hallucinated” case law and generate false citations, in the courtroom.

Attorneys in Maryland have been referred to disciplinary proceedings due to fake citations, while some argue that the technology has given pro se litigants unprecedented access to courts.

In federal court, there are no rules expressly banning AI from being used to draft legal briefs. But federal rules do require that attorneys, as well as pro se litigants, conduct a reasonable inquiry into the basis for the legal arguments and factual assertions made in any filings.

Many AI-related orders that have come down from judges have stemmed from “hallucinated” case citations — either to “a case that doesn’t exist, or a citation to a case that does not actually support the position for which it’s being cited,” said Parker E. Thoeni, managing shareholder of Ogletree Deakins’ Baltimore office.

Thoeni also said that pro se litigants using AI has “become much more prevalent than it was, even a year ago.”

“Everyone who’s in court has a responsibility to ensure that what they are providing to the court is actually accurate,” said Thoeni. He added that attorneys should be “wary of using AI to wholly create something, rather than as a tool to get something started or assist with review.”

James Rubinowitz, a New York-based trial attorney and lecturer at the Cardozo School of Law, said, “Because there has been just so many situations now where attorneys have made these filings with AI slop, we’re now relying on court precedent in order to see how lawyers should be governed.”

He added that the rules give broad discretion to judges as far as how to punish litigants who violate their duty of candor.

“It’s really the wild west right now with AI,” Rubinowitz said. “It’s not regulated, and each judge at this point has a different standard to what they hold these bad actors to.”

Abelson’s order on Monday stemmed from a case in which two brothers, Ali Behroz Aziz and Shinkay Aziz, sued their younger sibling and his attorney, as well as the U.S. and the Maryland attorneys general’s offices and numerous Montgomery County officials. Among other allegations, the brothers say they were illegally coerced into settling a prior case against their younger sibling. The plaintiff brothers and their younger sibling did not immediately return requests for comment Tuesday.

The “flurry of lengthy filings” cited by Abelson followed the judge’s order for the plaintiffs to show cause for why their case should not be dismissed.

“If Plaintiffs continue to file frivolous motions or include fabricated citations, and/or violate the page limitation set forth in the accompanying order, the Court may enjoin them from submitting further filings or consider imposing other sanctions,” he wrote.

This story has been updated with Rubinowitz’s quotes.