Judge Threatens Sanctions as Lawyers Defend AI-Written Filings in Voting Case

A federal judge in Texas has sharply criticized three attorneys for submitting court filings that drew on artificial intelligence, calling the work lazy lawyering and warning that sanctions may follow. US District Judge Mark Pittman made the remarks in a November 4 order rejecting a request to release Khalwinder Singh Bhengura, a 70-year-old Indian national and lawful permanent US resident, from custody before sentencing. The dispute became public in reports published on November 7 and November 8.

The Order and the Case

Bhengura was convicted on October 27 of falsely claiming American citizenship in order to register to vote and of illegally voting in the 2024 federal election. His sentencing is scheduled for February 11, 2027. His lawyers, Mario Herrera, Nicole Knox and Vincent Rene Carrizales, filed a motion seeking his immediate release, arguing that sentencing guidelines recommending zero to six months of imprisonment justified ending his pre-sentence detention. The motion acknowledged that portions of it had been prepared with the assistance of a generative artificial intelligence tool.

Pittman rejected the reasoning and described the defense reading of the relevant statute as abysmal. He called one citation an example of the risks lawyers take when they rely on artificial intelligence to draft briefs, and he expressed sympathy for prosecutors who had to answer what he termed defense by ChatGPT. The judge said the court was exhausted by the filings and questioned the lawyers' professional competence.

Sanctions and the Defense Response

The order signaled that disciplinary action was warranted, potentially including restrictions on the attorneys' ability to practice law in the Northern District of Texas. Pittman said scheduling constraints would prevent a hearing before the Thanksgiving holiday. That timing leaves the question of penalties open while the underlying case proceeds toward sentencing.

The three attorneys rejected the criticism. Knox said the team had nothing to hide and argued that its legal strategy was original and could not have been produced by an AI tool. Herrera said he had personally written, reviewed and finalized the filings and that none of the issues raised were frivolous, adding that he saw no violation of court rules. The exchange places a routine question of legal drafting before a judge who has made his view of automated assistance clear.

For lawyers and the staff who support them, the case touches on legal compliance in everyday practice: who reviews machine-assisted text, what disclosures a filing should carry, and how quickly a lawyer must be able to defend every line submitted to a court. Firms have increasingly used such tools for case-law research, organizing legal issues and preparing early drafts, work that junior lawyers have traditionally handled. The judge's order shows that courts can scrutinize the final product closely, regardless of how it was produced.

What to Watch Next

The next steps are the court's decision on whether to impose sanctions and the sentencing now set for February. Both will be watched by practitioners weighing how far automated drafting can go before a judge treats it as a failure of professional judgment. The case also puts attention on review routines inside law offices, where senior attorneys are expected to check phrasing and legal reasoning before anything reaches a court.