Legislation mandates disclosures for high-risk AI system deployments starting October 2026

On April 9, 2025, the Connecticut State Legislature introduced Senate Bill 2, a significant piece of legislation aimed at regulating high-risk artificial intelligence (AI) systems. The bill seeks to address growing concerns over the ethical deployment of AI technologies, particularly in preventing algorithmic discrimination and ensuring transparency in AI operations.

The main provisions of Senate Bill 2 include requirements for deployers of high-risk AI systems to enter into contracts with developers, ensuring that developers assume certain responsibilities related to the AI's deployment. Notably, deployers must not rely solely on their own data for training these systems, promoting a broader data usage approach. Additionally, the bill mandates that deployers provide consumers with impact assessments completed by developers, which must include detailed information about the AI's performance and potential risks.

A critical aspect of the bill is its focus on accountability. If a deployer discovers that their AI system has caused algorithmic discrimination affecting at least 1,000 consumers, they are required to notify the Attorney General within 90 days. This provision aims to enhance oversight and prompt action against discriminatory practices in AI applications.

The introduction of Senate Bill 2 has sparked notable debates among legislators and stakeholders. Proponents argue that the bill is essential for protecting consumers and fostering ethical AI development, while opponents raise concerns about the potential burden on businesses and the feasibility of compliance. Amendments to the bill have been proposed to address these concerns, but discussions remain ongoing.

The implications of Senate Bill 2 are significant, as it positions Connecticut as a leader in AI regulation, potentially influencing similar legislative efforts in other states. Experts suggest that the bill could set a precedent for balancing innovation with consumer protection, although its success will depend on effective implementation and enforcement.

As the legislative process unfolds, stakeholders will be closely monitoring the bill's progress and its potential impact on the rapidly evolving landscape of artificial intelligence. The next steps will involve further discussions and possible revisions before a final vote is scheduled.