Connecticut’s SB5 passed both chambers on May 1 and now heads to the governor, with Governor Ned Lamont confirming he will sign it. The bill cleared the House by 131-17 and the Senate by 32-4, drawing bipartisan support in each chamber. Formally titled the Connecticut Artificial Intelligence Responsibility and Transparency Act, the measure is shaping up as one of the broadest state AI laws in the US.
SB5 reaches hiring systems, AI companions, and synthetic media
The law applies to AI companions, automated employment decision tools, synthetic media provenance, and frontier model developers that meet defined thresholds. Its first effective date is October 1, 2026. Most provisions will be enforced only by the state Attorney General under unfair or deceptive trade practices law, and the bill creates no private right of action.
For employers, SB5 requires disclosure when automated tools are used in recruiting or hiring decisions. It also blocks companies from using those tools as a defense against discrimination claims. Those employment-related rules take effect on October 1, 2026.
Rules for AI companions, aimed at chatbots that encourage emotional attachment, are scheduled to take effect in January 2027. Generative AI systems with more than 1 million users must adopt provenance data standards aligned with C2PA.
Frontier developers face internal safety obligations
SB5 also requires frontier model developers to maintain internal AI safety programs and protect employees who report safety concerns. The report noted that regulation around companion AI has accelerated across US states in 2026 after lawsuits in Pennsylvania and Kentucky tied to alleged chatbot harm.
The bill also creates a regulatory sandbox and a working group. Its first meeting must be held by August 31, 2026, setting up the implementation process. For companies operating across state lines, that adds another layer to an enforcement picture already shaped by federal agencies deploying AI tools to address regulatory gaps.
State lawmakers move ahead despite federal pressure
Connecticut is joining California, Colorado, and other states that have enacted AI-focused laws even after an executive order from the Trump administration urged states to avoid rules viewed as overly burdensome. The White House said that order was meant to preempt certain state measures, but Connecticut moved ahead with its own framework.
Attorney General William Tong said in a February 2026 advisory to businesses that his office already viewed AI as clearly within its remit. SB5 now gives that office a broader and more purpose-built set of enforcement tools.