Connecticut Enacts Omnibus AI Law With Frontier-Model Whistleblower Shield and Provenance Mandate

Connecticut Governor Ned Lamont signed Senate Bill 5 into law on May 27, giving the state one of the nation's most comprehensive AI statutes. The 39-section omnibus measure imposes obligations on consumer-facing AI chatbots, frontier AI model developers, automated employment decision tools and social media platforms.

Its frontier provisions cut closest to the model labs: developers training foundation models using more than 10^26 computing operations must protect employees who report concerns that the developer or a model may contribute to a 'catastrophic risk,' defined as an event causing injury or death to 50 or more people or $1 billion-plus in property damage from CBRN assistance, autonomous cyberattacks or autonomous criminal conduct. Large generative AI providers — those with more than one million monthly users — must embed provenance data into any audio, image or video content their systems generate or materially alter.

Employers must disclose use of an 'automated employment-related decision process' to applicants and workers, and companion-chatbot operators face safety protocols with extra measures for minor users plus consumer disclosures for subscription AI products. Most provisions take effect October 1, 2026; AI-companion requirements begin January 1, 2027, and employment-tool obligations apply to systems deployed on or after October 1, 2027.