Developers required to disclose algorithmic discrimination risks in AI systems by 2026

On April 9, 2025, the Connecticut State Legislature introduced Senate Bill 2, a significant legislative proposal aimed at regulating high-risk artificial intelligence (AI) systems. The bill seeks to address growing concerns over algorithmic discrimination and the ethical implications of AI technologies in various sectors.

The primary purpose of Senate Bill 2 is to establish a framework for developers of high-risk AI systems, mandating them to disclose known risks associated with algorithmic discrimination. This disclosure is required to be made to the Attorney General and to all known deployers or other developers involved with the AI system. The bill stipulates that developers must report any instances of algorithmic discrimination affecting at least one thousand consumers, either discovered through their testing or reported by deployers, within a strict timeline of ninety days.

Key provisions of the bill include the obligation for multiple developers contributing to a high-risk AI system to adhere to the same disclosure requirements, albeit limited to their specific contributions. Importantly, the bill also includes protections for trade secrets and information that could pose security risks, ensuring that developers are not compelled to disclose sensitive data.

The introduction of Senate Bill 2 has sparked notable debates among lawmakers and stakeholders. Proponents argue that the bill is a necessary step toward accountability in the rapidly evolving AI landscape, emphasizing the need for transparency to protect consumers from potential biases embedded in AI algorithms. Critics, however, raise concerns about the feasibility of compliance for developers, particularly smaller firms that may lack the resources to navigate the regulatory landscape.

The implications of this bill extend beyond technical compliance; it touches on broader social and ethical issues surrounding AI technology. Experts suggest that if passed, Senate Bill 2 could set a precedent for similar regulations in other states, potentially influencing national standards for AI governance. The bill's focus on algorithmic discrimination aligns with increasing public scrutiny of AI systems and their impact on marginalized communities.

As the legislative process unfolds, the future of Senate Bill 2 remains uncertain. Lawmakers will continue to debate its provisions, and amendments may be proposed to address concerns raised during discussions. The outcome of this bill could significantly shape the regulatory environment for AI technologies in Connecticut and beyond, marking a pivotal moment in the intersection of technology and public policy.