Should an AI brief contain factitious citations, a fine can follow

On September twenty-fourth, CourtListener RECAP dockets: generative AI and sanctions or hallucinations reported that, in Clark v. Social Security, Commissioner of, a court ordered Rich to show cause why she should not be sanctioned, including with the imposition of a fine, for including factitious citations in a plaintiff’s brief. The filing says generative artificial intelligence was used in preparing the brief and records Rich’s response to the show cause order. The order therefore moves the matter beyond a warning about drafting technology and into a formal sanctions process.

The order exemplifies the basic operating fact for AI-native legal work: AI does not eliminate the consequences of bad legal work. A generated brief remains a filing, and factitious citations remain a defect that can trigger a court response. The important point is not that AI was used. The important point is that the court connected the filing’s citations to a show-cause process that may include a fine. That is accountability attached to the finished legal product, not to the novelty of the tool used to produce it.

The Clark order proves that a court can investigate sanctions where an AI-assisted brief contains factitious citations. It does not prove that a fine has been imposed, that every AI-assisted filing is unreliable, or that the technology itself caused the citations. The court’s next decision matters because a show-cause order is a procedural step, not the final consequence. A separate CourtListener: AI and privilege (opinions) entry describes Douglas v. Deutsche Bank National Trust Co., Published Order as a cautionary tale about the misuse of AI and its consequences for attorneys, showing that courts can frame misuse as a professional consequence without establishing a general rule for every legal filing. Macias v. CCADC supplies a narrower comparison: its docket says OpenAI Codex assisted in preparing a document and that the signer certified personal checking of the AI-assisted portions, including all citations. That certification does not establish accuracy, but it identifies the verification step required for a stronger claim that AI can support finished legal work. Clark supplies no comparable finding that the citations were checked successfully.

The next observable fact is the court’s ruling on the Clark show-cause order: whether it imposes a fine or another sanction, declines to sanction Rich, or identifies a different failure in the brief. That order will determine whether the immediate consequence remains a formal warning process or becomes a monetary or other penalty. A later filing that documents reliable personal review of every citation would support the narrower case for AI-assisted legal work. A ruling that sanctions the filing would reinforce the broader lesson: finished legal work still has to survive verification, and AI does not remove the consequences of bad legal work.