Requires disclosure of use of generative artificial intelligence to clients, criminal defendants, and the court.
A09097 would require disclosure when generative artificial intelligence is used in drafting legal papers filed in New York courts. The bill defines “generative artificial intelligence” broadly to include machine learning systems, software, automation, algorithms, and related technologies used for research, document review, and document creation. It would add a new Civil Practice Law and Rules provision for civil cases and a new Criminal Procedure Law section for criminal cases, with parallel requirements in both contexts.
Under the bill, courts would have to notify counsel and self-represented litigants, or in criminal cases counsel and pro se defendants, about the risks of using generative AI for legal work and about the new rule’s requirements. A paper or file drafted with AI could not be submitted unless the client or defendant gave informed consent after being warned of those risks. Any filing prepared with AI assistance would need a separate affidavit disclosing that use and certifying that a human reviewed the source material and verified the accuracy of the content, including citations. If AI was not used, a separate affidavit stating that fact would be required. The bill also amends appellate briefing rules to require an AI-use disclosure and human-verification certification when applicable.
The bill’s main legal impact would be to create a new disclosure and certification regime for attorneys and litigants in both civil and criminal proceedings, affecting court filings statewide. It would not ban generative AI in legal practice, but it would impose notice, consent, and affidavit requirements that could change how lawyers, public defenders, and self-represented parties prepare pleadings, briefs, and other court papers. The act would take effect 90 days after becoming law.
The available context shows no committee transcript or recorded votes, so there is no documented floor or committee sentiment to assess. Based on the bill text and caption, the measure appears aimed at transparency and accuracy concerns surrounding AI-assisted legal drafting, with an emphasis on preventing unreviewed or unreliable AI-generated content from entering the court record.
The main point of contention likely would be the burden the bill places on attorneys, courts, and pro se litigants, especially the requirement for separate affidavits in every filing and the broad definition of generative AI. Supporters would likely emphasize client protection, accuracy, and accountability, while critics may argue that the disclosure rules are administratively heavy, may chill useful AI adoption, and could be difficult to apply consistently across routine legal work.
The bill would amend the Civil Practice Law and Rules and the Criminal Procedure Law to add new statewide requirements governing AI-assisted legal drafting. It would require courts to provide notice about generative AI risks, require informed client or defendant consent before AI-assisted drafting, and mandate disclosure affidavits certifying human review and accuracy verification. It would also modify appellate practice rules to require an AI-use disclosure when applicable, affecting attorneys, litigants, public defenders, and self-represented parties who submit papers in New York courts.
No committee discussion or vote history is provided, so there is no direct evidence of legislative support or opposition. From the bill’s structure and caption, the measure appears motivated by concern over the reliability and transparency of generative AI in legal filings, suggesting a generally cautious or regulatory approach rather than an anti-technology stance. The bill’s tone is protective and compliance-focused, aiming to ensure human oversight and informed consent.
The likely areas of contention are the breadth of the definition of “generative artificial intelligence,” the requirement that courts warn all counsel and pro se parties, and the mandate for separate affidavits in filings whether AI was used or not. Opponents may view these requirements as duplicative, burdensome, and potentially confusing for litigants, especially self-represented parties. Supporters are likely to argue that the disclosure and verification rules are necessary to prevent inaccurate citations, hallucinated content, and undisclosed AI reliance in court submissions.