The proposal follows a pattern of high-profile errors. Law firms have submitted briefs containing citations to nonexistent cases and fabricated legal quotations, prompting the Rules Committee to act. The U.S. District Court for Connecticut already issued a no-tolerance policy in September 2025 warning that unverified AI output violates Federal Rule of Civil Procedure 11. Connecticut's Bar Association and the Judicial Branch's Committee on Artificial Intelligence are studying the issue but have not yet issued binding guidance. The Rules Committee has set a public comment deadline of April 29, 2026, and scheduled a hearing for May 4, 2026.
Attorneys in Connecticut should prepare for imminent change. The proposed rule will likely become binding and will shift verification responsibility entirely to the filer. Any AI-generated content—whether a citation, a legal principle, or supporting evidence—must be independently confirmed before submission. This standard aligns with federal district court expectations and will likely influence how other states approach AI regulation. The rule's adoption would establish a clear liability framework for AI-assisted legal work and eliminate any ambiguity about attorney responsibility for machine-generated errors.