Should AI draft a brief, citation review remains counsel’s burden

On September 24, 2026, a court in Clark v. Social Security, Commissioner of ordered Rich to show cause why she should not be sanctioned, including with a fine, for including factitious citations in a plaintiff’s brief. CourtListener RECAP dockets: generative AI and sanctions or hallucinations reports that the brief used generative artificial intelligence and that the court issued the order after that use. The order does not merely identify an AI-assisted filing; it connects the filing to a judicial demand for an explanation and possible financial punishment.

The order moves the question from whether AI can produce legal text to who answers when the text contains false citations. AI does not eliminate the consequences of bad legal work. The consequence here is not yet a final sanction, but the court has made review of the finished brief part of counsel’s immediate problem. CourtListener: AI and sanctions (opinions) describes Douglas v. Deutsche Bank National Trust Co. as a cautionary tale about the misuse of AI and its consequences for attorneys. CourtListener: AI and privilege (opinions) reports that Cervantes v. Bianco disapproved improper use of AI tools where that use apparently contributed to the underlying problem. Together, the orders and opinions make accountability the limiting condition on AI-assisted legal work, rather than a feature that disappears when software prepares the words.

What Clark proves is narrower and more useful: a court ordered a lawyer to explain an AI-assisted brief after identifying factitious citations, and it left a fine among the possible outcomes. It does not establish that Rich will ultimately be sanctioned, that every AI-assisted brief contains citation errors, or that an AI-native law firm has failed. The stronger reading would require a final order imposing sanctions and explaining the review failure that justified them. It would also require more than a single matter to show a repeatable accountability pattern across finished legal work. The decision therefore supports a concrete proposition: AI use does not transfer responsibility for citations away from the person submitting the brief. A filing in Macias v. CCADC illustrates the standard that would matter in a different direction, because it includes an AI disclosure and a certification of personal checking of AI-assisted portions, including citations. That certification is a process statement, not proof that the citations were correct.

The next fact to watch is the final order in Clark v. Social Security, Commissioner of: whether the court imposes a sanction, whether it includes a fine, and what the court says about review of the citations before filing. That order would show whether the show-cause process became a consequence or remained a warning. A result that ties the sanction to the failure to verify citations would strengthen the accountability reading. A decision declining sanctions after crediting a documented review process would narrow it without eliminating the underlying burden. We expect the decisive measure to be the court’s treatment of the submitted work, not the presence of AI in the drafting process.