MyPillow Attorneys Filed an AI-Drafted Brief Full of Errors and Hid It from Their Client
What happened
In February 2025, two attorneys representing MyPillow CEO Mike Lindell submitted a court filing riddled with errors, the kind of mistakes that draw sanctions, judicial rebuke, and direct accusations of incompetence. The document had been drafted using an AI tool. Neither lawyer had adequately verified what the system produced before filing it in an active legal proceeding.
The court caught the errors. The attorneys did not volunteer them. Only after the court's intervention did the lawyers acknowledge the problems. Their explanation was that they had accidentally submitted an earlier draft, one in which AI-generated mistakes had not yet been corrected. That account raised more questions than it answered. A properly reviewed filing does not coexist with an unreviewed draft in a state that could be filed by accident.
The case was an election-related libel suit, a high-stakes proceeding with direct legal and financial consequences for Lindell. Despite that context, the attorneys did not inform Lindell that AI had been involved in preparing his defense. He learned about the AI use through the court proceedings, not from his own legal team. The lawyers exposed their client to risk while keeping from him the facts about how his defense was prepared.
The sanctions the court issued addressed the specific filing, but the conduct those sanctions punished was a pattern, not an isolated mistake. Courts have already seen the same failure mode play out in multiple proceedings: an attorney files what an AI system produced without independently verifying whether the citations exist, whether the claims are accurate, or whether the legal reasoning holds. The New York lawyer who cited fabricated case law and the defense attorney whose AI-assisted closing argument collapsed in court are separate incidents, but they follow the same path from model output to filed document with no confirmed review in between.
What these cases share, beyond the specific errors, is the absence of any mechanism that forces verification into the record before a document is filed. A provable record of what a system generated, which claims a human reviewed, and what was confirmed accurate before signing would change the accountability structure of legal AI use entirely. Without it, verification remains a professional norm enforced only after errors surface in court, which means the judiciary absorbs the cost of catching what the attorney should have confirmed at the desk.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.