The Connecticut Supreme Court waded into new territory in a hearing last week after AI hallucinations appeared in briefing before the high court. Associate Justice Steve Ecker, co-chair of the Committee on Artificial Intelligence in the Connecticut Legal System, said a motion for sanctions against plaintiff's counsel in a landlord-tenant dispute is the first time the court had to grapple with the issue. "The problem with AI," Ecker said, "we just don't know what's true anymore. It further plunges us into this world where it's very difficult to distinguish fact from fiction." Lawyers representing landlord Kosel Equity—Ian H. Gottlieb, David E. Rosenberg and Paul J. Small of GLG Law—admitted to using generative AI through platforms Lexis for legal research and ChatGPT to organize and format the content, however, the AI created errors in citations and references, according to court documents. The high court ordered the plaintiff to file a memorandum explaining how and when the errors occurred. In the memorandum, the attorneys stated that they "failed to properly proof its citations and references" during the review process. When the plaintiff's attorneys discovered the error on Feb. 10, they reached out to the chief clerk on Feb. 11 to self report and let the court know. Gottlieb did not respond tot a request for comment. The memorandum said the AI tools were not used as a substitute for legal research but a tool after initial drafting. "AI was used to assist in the organization and formatting of the brief, specifically with analyzing the brief to avoid duplication of arguments," the document said. "AI was also used to assist in reviewing the content of the brief in particular to comply with the word count restrictions. The errors identified in the errata sheet were corrected by manually checking the brief’s quotations and formatting against the underlying sources. Unfortunately, counsel did not notice that AI had intuitively made changes to the brief prior to filing." The memorandum called the situation "professionally embarrassing" and apologized to the court, the other parties and their client. "Counsel takes full responsibility for these errors and recognizes that they fell short of the standards expected in this court as well as in the legal community, but wishes to make clear that this was in no way intended to willfully violate any ethical obligation that counsel has to the court," plaintiff's counsel said in the memorandum. "Counsel wishes to assure the court that this is not an error that will happen again" The high court ordered the plaintiff to file a corrected brief. The Connecticut District Court issued a notice in September, stating it had a no-tolerance policy for AI hallucinations, and filing documents with such errors could lead to sanctions on the state court level. The Rules Committee of the Superior Court also has addressed proposed guidelines, such as requiring attorneys to append a certification to their pleadings or briefs, expressing they have verified the accuracy of the content provided by AI programs to conduct legal research. On Feb. 9, the Rules Committee again discussed the proposed requirement, taking into consideration comments and responses from judges, attorneys, and lawyer associations. Counsel for the defendant in the case, Philip G. Kent of Brenner, Saltzman & Wallman, declined to comment on the matter. Read the memorandum: