California's State Bar says three lawyers leaned on generative AI for their legal research and ended up filing briefs laced with nonexistent or off-point case citations. Two of those attorneys now face formal disciplinary charges, while a third has accepted conditions that include a short suspension, as regulators tighten the screws on machine-assisted lawyering.
According to the Los Angeles Times, the Office of Chief Trial Counsel has lodged notices of disciplinary charges against Omid Emile Khalifeh and Steven Thomas Romeyn, and the State Bar Court has signed off on a disciplinary stipulation for Sepideh Ardestani. The Times reports that the State Bar Court has not yet ruled on the allegations involving Khalifeh and Romeyn, while Ardestani agreed to probation and a brief suspension under a stipulation the court approved on April 6.
California courts have repeatedly signaled that generative AI may be a drafting aid, not a replacement for a lawyer’s duty to check the law and the facts. The Supreme Court of California has cautioned that AI tools can "hallucinate" legal authorities and has urged attorneys to verify any machine-generated research before it ever hits the court’s docket.
The charges and the stipulation
As outlined by the Los Angeles Times, Khalifeh faces six counts of alleged misconduct tied to an April 2025 federal trademark filing in which the State Bar contends he cited one case that did not exist, relied on two citations that did not actually support his arguments, and ran afoul of a standing order that required him to disclose any use of generative AI. Romeyn is accused of filing an October 2025 brief in Orange County that included several citations that were either made up or irrelevant, and he has acknowledged that he did not independently confirm each citation before submitting the document.
The State Bar Court has approved a stipulation for Sepideh Ardestani that calls for one year of probation, a 30-day suspension, and ten hours of continuing legal education focused on technology. Under the State Bar’s procedures, that recommendation now heads to the California Supreme Court for final review, according to the State Bar of California.
Judges are already cracking down
Judges in California and around the country have started backing up their warnings with penalties as AI-generated phantom cases show up more often in court filings. Sanctions have included fines, orders striking tainted briefs, and sharp words about how fabricated citations burn scarce judicial time. Law firms and monitors such as Ropes & Gray are tracking a growing list of AI-focused court orders, and recent reporting has highlighted multiple matters where judges concluded that AI-driven missteps had wasted limited court resources.
What to watch next
The State Bar Court will decide whether the pending notices of disciplinary charges justify formal discipline for Khalifeh and Romeyn, and any recommendation could go to the California Supreme Court, which has the power to impose suspension or disbarment. The situation is likely to intensify calls for mandatory AI training and clearer disclosure rules. At the same time, the State Bar’s public guidance keeps stressing a basic point that predates any chatbot: lawyers are accountable for the accuracy of everything they file, even when an algorithm helps draft it, according to the State Bar of California.
