Connecticut’s SB5 legislation cleared both legislative chambers on May 1 and is now awaiting the governor’s signature, positioning it among the most far-reaching state-level AI laws in the United States.
The bill passed with overwhelming bipartisan backing, securing a 131-17 vote in the House and a 32-4 vote in the Senate. Governor Ned Lamont has already indicated that he intends to approve the measure, officially known as the Connecticut Artificial Intelligence Responsibility and Transparency Act.
The legislation introduces regulations covering AI companions, synthetic media disclosure standards, automated hiring and recruitment systems, as well as advanced frontier AI developers that exceed certain operational thresholds.
Its first provisions will come into force on October 1, 2026. Most enforcement authority will rest solely with the state Attorney General under unfair and deceptive trade practice laws, while the bill does not provide individuals with a private right to sue.
Under SB5, employers will be required to inform applicants and workers whenever automated AI systems are involved in recruitment or employment decisions. Companies will also be prohibited from citing AI tools as a defense in discrimination-related claims. These employment-focused requirements are scheduled to begin in October 2026.
Rules targeting AI companion platforms — particularly chatbots designed to build emotional connections with users — are set to take effect in January 2027. Additionally, generative AI platforms with more than one million users must implement provenance and authenticity standards aligned with the C2PA framework.
The legislation also mandates that frontier AI developers create internal safety and governance programs while offering protections to employees who raise concerns about AI-related risks or misconduct.
Connecticut’s move comes despite efforts by the Trump administration to discourage states from introducing what it considers restrictive AI regulations. With SB5, Connecticut joins states like California and Colorado in pushing forward with independent AI governance frameworks.
The law further establishes a regulatory sandbox and a dedicated working group tasked with shaping implementation guidelines, with the group’s first meeting required by August 31, 2026.
Connecticut Attorney General William Tong previously signaled that his office already considered artificial intelligence to fall under existing consumer protection oversight. SB5 significantly expands the legal tools available to his office for AI-related enforcement actions.
