The State Bar Is Done Warning People

The State Bar of California has filed formal disciplinary charges against two attorneys — Omid Emile Khalifeh of Los Angeles and Steven Thomas Romeyn of Scottsdale, Arizona — for submitting AI-generated court filings that cited nonexistent legal decisions, according to the Los Angeles Times. A third attorney, Sepideh Ardestani of Beverly Hills, has already had disciplinary measures approved by the State Bar Court after she submitted fake and erroneous citations in a March 2025 federal court filing.

Six separate misconduct charges have been filed against Khalifeh alone, according to the LA Times.

None of the three attorneys could be reached for comment.

$10,000 Fine. 21 Fake Cases. One ChatGPT Session.

California attorney Amir Mostafavi was hit with a $10,000 fine after filing an appeal that contained 21 fabricated case citations generated by ChatGPT, according to the Associated Press as reported by The Daily Record.

The court made clear that lawyers must personally verify every citation. The case involved a straightforward mandate: verify citations rather than relying on the AI's output.

A Judge Actually Ruled on a Fake Case

In H.C. v. Contreras, a California Court of Appeal — with Justice Mark Snauffer writing the opinion, joined by Justices Bert Levy and Donald Franson — reversed a trial court ruling because the trial judge cited and relied on a fictitious case submitted by the opposing attorney's counsel, according to Reason's Volokh Conspiracy blog.

The case involved a mother, Bethany G., seeking a protective order for her minor son against the child's father, Rudy C. Rudy's attorney submitted a closing brief citing Enrique M. v. Angelina V. (2005) 15 Cal.App.5th 788 — a case that does NOT exist.

Bethany's counsel caught it immediately and flagged it in writing.

The trial court ignored the warning and ruled anyway — using the fake case as part of its reasoning to deny the protective order.

A real family, a real child, a real safety hearing — and the outcome rested partly on a case that was never decided, never published, and never existed. The appeals court reversed the ruling and sent it back for new proceedings.

What Mainstream Coverage Is Getting Wrong

Most media outlets are framing this as a story about AI being dangerous or lawyers being careless.

The actual story involves a two-layer failure.

First, attorneys are treating AI tools like paralegals with law degrees. ChatGPT is a text prediction engine. It doesn't know what's true — it knows what sounds plausible. Second, courts are letting it happen. The trial judge in H.C. v. Contreras had a written warning from opposing counsel sitting in front of them. They ruled on the fake case anyway.

The LA Times piece does solid work naming names, which is more than most outlets bother to do. But it buries the H.C. v. Contreras appellate reversal — the most concrete proof of real-world harm — entirely. A child's protective order got denied because of a hallucinated citation.

The California Judicial Council Is Moving

The California Judicial Council is now moving to formally regulate AI use in court filings, according to The Daily Record. Details on the specific rules are still developing, but mandatory disclosure and verification requirements are coming.

State Bar chief trial counsel George Cardona put it plainly in a statement to the LA Times: "Courts and clients must be able to trust that the filings attorneys submit are accurate, supported, and compliant with professional standards. Technology can assist legal practice, but it does not replace an attorney's duty of competence, diligence, and honesty."

What This Means for Regular People

If you're hiring a lawyer, ask them directly whether they use AI to draft filings and what their verification process looks like. You have every right to know.

If they can't answer clearly, that's significant.

Attorneys are supposed to be officers of the court. That title carries an obligation to tell the truth, cite real law, and do the basic homework. A $10,000 fine and a disciplinary charge are the baseline response to submitting 21 fake cases or getting a child's protective order denied on invented legal authority.

The California bar is finally holding attorneys accountable. The courts need to match that energy — starting with sanctions on judges who rule on citations after being warned in writing that those citations don't exist.