Establishes regulations to ensure the ethical development, integration, and deployment of high-risk AI systems, particularly those influencing consequential decisions.
S0627 creates a new chapter in Rhode Island commercial law titled the “Artificial Intelligence Act” to regulate the development, integration, and deployment of high-risk artificial intelligence systems. The bill focuses on AI used in consequential decisions affecting employment, education, lending, housing, insurance, legal services, essential government services, and healthcare. It defines key terms such as high-risk AI system, deployer, developer, integrator, algorithmic discrimination, synthetic digital content, and general-purpose AI model, and sets out compliance obligations beginning October 1, 2026.
The bill requires developers, integrators, and deployers to use reasonable care to prevent algorithmic discrimination, maintain risk management policies, conduct impact assessments, provide documentation and public-facing disclosures, and notify consumers when AI is used in consequential decision-making. It also requires certain adverse decision notices, consumer access to reasons and data, and opportunities to correct inaccurate data and appeal decisions in some cases. In addition, it mandates that AI-generated synthetic digital content be marked and detectable, subject to exceptions for certain text, artistic, or low-risk uses.
The bill would add a comprehensive AI regulatory framework to Title 6 of the Rhode Island General Laws and create new duties for businesses that develop, sell, integrate, or use high-risk AI systems in the state. It would also authorize the attorney general to enforce the chapter exclusively, treat violations as unfair trade practices, and request documentation during investigations, while preserving trade secret protections and attorney-client/work-product privileges. The act includes exemptions for certain federally regulated systems, some financial institutions and insurers subject to equivalent oversight, internal-use systems, and other specified categories.
Based on the bill text and explanation, the overall sentiment appears supportive of AI oversight and consumer protection, with the bill framed as promoting ethical development and transparency rather than banning AI use. The legislation is structured to encourage compliance through risk management, disclosure, and a cure period before enforcement, suggesting a regulatory approach intended to be workable for industry. No committee transcript or vote record was provided, so there is no recorded opposition or support from hearings or floor action in the supplied materials.
The main points of potential contention are the scope and burden of compliance for AI developers, integrators, and deployers, especially the requirements for impact assessments, public disclosures, consumer notices, and ongoing monitoring. Another likely issue is the breadth of the bill’s definitions of high-risk AI and consequential decisions, which reach into employment, housing, lending, education, healthcare, and legal services. Industry stakeholders may also object to reporting obligations and attorney general access to documentation, while proponents are likely to emphasize the bill’s trade secret protections, exemptions for federally regulated systems, and carveouts for internal-use or low-risk applications.