Posted by Meaghan Kirby in Artificial Intelligence, Connecticut Law
Jul 21 2026
Connecticut’s courts have adopted new rules governing the use of generative artificial intelligence in court filings. The rules make one principle clear: although lawyers and self-represented parties may use AI tools to assist with legal work, the person filing a document remains fully responsible for its accuracy. The changes follow increasing concerns about AI-generated briefs containing nonexistent cases, inaccurate quotations, and other seemingly authoritative errors.
Effective June 23, 2026, new Connecticut Practice Book Section 4-9 applies whenever generative AI is used to create or edit a document filed in Connecticut Superior Court. The filer must independently verify every citation, legal authority, and other item produced by the AI tool. By submitting the document, the filer represents that they have reviewed the rule and made good-faith, diligent efforts to ensure compliance with their legal obligations, including under the Rules of Professional Conduct governing the practice of law in the state. The responsibility rests solely with the person filing the document; it cannot be shifted to the technology or the court.
Similarly, the Connecticut Supreme and Appellate Courts recently adopted corresponding requirements. Under revised Rule of Appellate Procedure 62-6, when a person signs any paper (motion, brief, or otherwise), that signature represents that they personally read it, believes that there are good grounds to support it, and complied with the rules (including Practice Book § 4-9) governing generative AI. Revised Rule 85-2 permits the imposition of sanctions when a filer fails to independently verify legal citations and authorities produced by AI. These appellate rules took effect immediately upon publication in July 2026, following two cases in which an attorney acknowledged that three appellate filings contained ChatGPT-generated misquotations and citation errors because he had failed to review the final documents after using the program to improve his work. The Connecticut Supreme Court is considering whether those errors warrant sanctions.
Connecticut’s federal judges have taken a similar approach. In September 2025, the U.S. District Court for the District of Connecticut issued a notice, warning counsel and litigants that using AI without verifying its accuracy implicates Federal Rule of Civil Procedure 11 (which rule requires lawyers to certify the truthfulness of their representations to the court) and announced a “no-tolerance” policy for filings that hallucinate or invent legal propositions or severely misstate the law. The District warned that such filings may result in the imposition of sanctions.
AI tools may help lawyers work more efficiently, but they are not substitutes for legal judgment, careful research, or proofreading. The AI-related requirements of our state’s courts reinforce lawyers’ existing professional obligations, including their duties of competence, confidentiality, supervision, candor, and honesty.
Posted by Meaghan Kirby in Artificial Intelligence, Connecticut Law
Tagged AI, connecticut, federal judges, generative artificial intelligence, law, litigants
