Another federal court has sanctioned lawyers for using generative AI in legal drafting without verifying the results. On July 7, 2025, a U.S. District Judge ordered two attorneys representing MyPillow CEO Mike Lindell to each pay $3,000 after they filed a brief riddled with nearly 30 “defective citations,” including non-existent cases and misstatements of law.

The brief was submitted in February in a defamation lawsuit brought by a former Dominion Voting Systems executive, who accused Lindell and his media platform of promoting false claims that the executive had rigged the 2020 election. A Colorado jury later found Lindell and his company liable for over $2 million in damages.

The judge faulted the attorneys for submitting the flawed filing and misleading the court about its origins. One of the attorneys initially denied using AI, later claiming the wrong version of a “thoroughly reviewed” brief had been filed while he was on vacation in Mexico. But the so-called “correct” draft bore the same errors as the filed version and included metadata showing it was edited after the court had already flagged the issue. The judge also noted the attorneys had filed similarly defective documents in another case.

Although sanctions did not reach the clients, they are not immune from the consequences and inevitably suffer a loss of credibility. Courts are paying close attention to AI use in legal practice, and errors stemming from unverified outputs can undermine a case. Lawyers who rely on AI without rigorous review do so at their peril, as do the clients who rely on them.