Connecticut court sanctions lawyer over AI hallucinations in filings
Connecticut's highest court issued its first sanctions decision addressing AI-generated hallucinations on July 31, 2026, after appellate briefs in two cases contained fabricated and altered citations produced by generative AI tools.
The Connecticut Supreme Court ruled in TOV Realty, LLC v. Suarez (SC 21183) and Kosel Equity, LLC v. MacGregor (SC 21184) that Attorney Ian G. Gottlieb, a partner at GLG Law in Wallingford, had submitted approximately seven erroneous citations as a result of relying on generative AI without adequate verification.
The Ruling and Sanctions
The court imposed sanctions including six hours of continuing legal education—with three hours specifically focused on generative AI—plus $1,000 donations from both Gottlieb and his firm to the Connecticut Bar Institute for generative-AI education.
According to Reuters reporting, the Connecticut Supreme Court described this as its first sanctions ruling addressing AI hallucinations. The court stated: "This order should serve as a reminder to all members of the legal profession to learn about the risks and limitations of the technologies they use in the practice of law because they remain personally responsible to make certain that all information submitted to the court is true and accurate."
Why This Matters for the Legal Profession
The ruling establishes concrete accountability for lawyers who deploy generative AI without verification safeguards. Rather than a large monetary fine, the court emphasized education and professional responsibility—signaling that the bar must understand both the technical failure modes of AI systems and their obligation to independently verify every citation and claim submitted to courts.
The decision arrives as generative AI tools like ChatGPT and Claude have proliferated in legal workflows. Hallucination—the tendency of large language models to generate plausible-sounding but fabricated information—remains a well-documented risk, particularly when these tools are asked to cite specific cases, statutes, or precedents without access to real-time databases.
The Connecticut Supreme Court's warning applies pressure on the legal profession at a critical moment: as AI-assisted legal research and brief-writing become routine, courts are establishing that the burden of verification rests entirely with the attorney, not the technology vendor.
Implications
Other state bars and federal courts are likely to monitor this ruling closely. The decision reinforces that AI-assisted work does not diminish attorney accountability—a principle that could shape disciplinary standards nationwide as generative AI becomes embedded in legal practice.