If AI-assisted briefing fails review, legal consequences can follow

On September 24, the court in Clark v. Social Security, Commissioner of ordered Rich to show cause as to why she should not be sanctioned, including with the imposition of a fine, after a brief prepared with generative artificial intelligence included factitious citations, according to the CourtListener RECAP docket. The docket says Rich responded to the show-cause order and acknowledged using AI in preparing the brief. The immediate event is therefore not a product launch or a promise about legal automation. It is a court process directed at a lawyer after an AI-assisted filing contained defective legal work.

That process puts the central accountability question in plain view: AI does not eliminate the consequences of bad legal work. The filing technology may have helped prepare the brief, but the court still examined the citations and required a response from the lawyer. The consequence remains attached to the submitted work, not erased by the tool used to produce it. The point is reinforced by other CourtListener records, though each concerns its own facts. In Douglas v. Deutsche Bank National Trust Co., a published order describes the matter as a cautionary tale about the misuse of artificial intelligence and its consequences for attorneys. In Cervantes v. Bianco, the court says it disapproves of the improper use of artificial intelligence tools that apparently contributed to the underlying problem. Together, the orders show that judicial scrutiny can reach AI-assisted legal work when the work fails basic reliability checks.

Clark establishes a narrower proposition than the strongest marketing version of AI-native legal work. It establishes that a court ordered a lawyer to explain why sanctions, including a fine, should not follow a brief containing factitious citations after generative AI was used in its preparation. It does not establish that AI caused every defective citation, that the lawyer lacked any review process, or that a particular AI system produces unreliable work in every setting. It also says nothing about an AI-native firm as an organisation, its internal controls, its client outcomes, or its economics. The Douglas and Cervantes orders add judicial concern about misuse, but they do not turn Clark into a general performance study. A stronger conclusion would require final sanctions orders, repeated matters tied to a defined workflow, and numbers showing whether review prevents defective filings at a materially higher rate than ordinary legal work. None of that follows from the show-cause order alone.

The next fact to watch is the court’s final disposition of the Clark show-cause process. An order imposing a fine or another sanction would confirm that the defective AI-assisted filing produced a completed legal consequence, rather than only an inquiry. An order declining sanctions would narrow the lesson while leaving the citation failure and the court’s demand for an explanation intact. Further published orders in Clark, Douglas, or Cervantes would show whether these matters are isolated responses or part of a continuing judicial pattern. For now, the observable claim is limited but important: the use of AI does not move responsibility away from the legal work delivered to the court.