Connecticut's House of Representatives voted 131-17 on Friday to pass Senate Bill 5, sending what is now the most comprehensive state-level AI regulation in America to Governor Ned Lamont's desk. Lamont has confirmed he will sign it.

The bill had previously passed the Senate 32-4 after hours of floor debate. The bipartisan supermajorities in both chambers reflect years of negotiation over legislation that state lawmakers have attempted and failed to pass in previous sessions.

64 Pages Covering Nearly Everything AI Touches

SB 5 spans 64 pages and 37 sections. It is not a narrow regulation targeting one specific AI application. It is a structural framework that touches nearly every dimension of how artificial intelligence intersects with commercial life in the state.

Key provisions include regulations on emotional companion chatbots, which have drawn intense scrutiny after reports of AI chatbots forming inappropriate relationships with minors. The bill also covers automated hiring pipelines, requiring transparency and impact assessments when AI systems are used to screen, rank, or reject job applicants.

Frontier model developers face new safety requirements. Synthetic content must be labeled. State agencies must follow new protocols for any AI-related tasks and programs. And the bill establishes a publicly funded AI training academy to build workforce capacity.

Industry Opposition Was Not Enough

The bill faced lobbying from industry groups. NetChoice, a trade group representing major tech companies, sent an eight-page letter arguing against passage, warning that "enacting SB 5 creates significant risk for Connecticut" and urging the state to wait for federal guidance first.

Senator James Maroney, the bill's champion, defended it as primarily focused on transparency and disclosure, noting that many of the requirements are taken from existing law. The bipartisan vote margins suggest his framing resonated: the bill was not treated as an anti-tech measure but as a governance framework.

Why Connecticut Matters for the Rest of the Country

State AI regulation has been a graveyard of ambitious proposals. Colorado passed a narrower AI discrimination law in 2024. California attempted comprehensive regulation and the governor vetoed it. Multiple states have introduced AI bills that stalled in committee. Connecticut is the first to pass both chambers with a bill this broad and have a governor ready to sign.

The companion chatbot provisions are particularly significant. No U.S. state has enacted regulations specifically targeting AI systems designed to simulate emotional relationships. Connecticut is writing new law in a space where the technology has outrun the legal framework entirely.

The bill's breadth is what separates it from everything that came before. Rather than targeting one AI risk, Connecticut chose to build a regulatory architecture that covers the full surface area of how AI intersects with employment, consumer protection, content integrity, government operations, and frontier model development. Whether that ambition survives contact with enforcement remains to be seen. But the template now exists, and other states will be reading Connecticut's 64 pages closely.