Substitute Bill No. 2 introduces comprehensive regulations for the development, deployment, and integration of high-risk artificial intelligence (AI) systems in Connecticut, effective October 1, 2026. The bill mandates that developers, integrators, and deployers of AI systems exercise reasonable care to prevent algorithmic discrimination, requiring them to conduct impact assessments, maintain documentation, and disclose known risks to the Attorney General and affected parties. It establishes a rebuttable presumption of reasonable care for those who comply with the provisions, while also allowing for exemptions in certain cases, such as when developers release models under open-source licenses. The bill emphasizes transparency by requiring deployers to inform consumers when they interact with AI systems and to provide clear information about the systems' purposes and decision-making processes.

Additionally, the bill creates a regulatory framework for an AI sandbox program, overseen by the Department of Economic and Community Development, to facilitate the safe testing of AI technologies. It establishes a working group to address consumer protection and labor implications of AI, modifies definitions related to AI, and expands the legal framework surrounding the unlawful dissemination of intimate images to include synthetic content. The legislation aims to enhance consumer protection, promote responsible AI use, and ensure that AI technologies are developed and deployed in a manner that is equitable and transparent, while also providing a structured process for enforcement by the Attorney General.

Statutes affected:
Committee Bill: 10-21l, 32-7p, 32-39e
GL Joint Favorable Substitute: 10-21l, 32-7p, 32-39e
File No. 603: 32-7p, 32-39e
JUD Joint Favorable: 10-21l, 32-7p, 32-39e
APP Joint Favorable: 10-21l, 32-7p, 32-39e