Regulates the development and use of certain artificial intelligence systems to prevent algorithmic discrimination; requires independent audits of high risk AI systems; provides for enforcement by the attorney general.
S01169, the “New York artificial intelligence act,” would create a new Article 8-A in the Civil Rights Law to regulate the development, sale, deployment, and use of certain artificial intelligence systems in New York. The bill focuses on “high-risk AI systems” used in consequential decisions affecting employment, education, housing, family planning, health care, financial services, law enforcement, and legal services. It defines algorithmic discrimination, sets out duties for developers and deployers, and requires notice to affected users, opportunities to opt out of automated decision-making in some cases, and access to human review and appeals.
The bill also requires independent third-party audits, recurring reporting to the attorney general, and the adoption of a documented risk management policy and program based on recognized AI risk frameworks. It prohibits social scoring systems, creates whistleblower protections for employees, and gives the attorney general authority to enforce the law through injunctions and civil penalties. It also creates a private right of action for harmed persons and amends the Executive Law to make violations of the new AI discrimination rules an unlawful discriminatory practice under state civil rights law.
The bill would significantly expand New York civil rights and consumer-protection law by adding AI-specific obligations to the Civil Rights Law and tying violations into the Executive Law’s anti-discrimination framework. Developers and deployers of covered AI systems would face new compliance duties, including audits, reporting, notice, human-review procedures, and risk-management documentation, while the attorney general would gain enforcement authority and public reporting oversight. The bill would also create potential civil liability, including compensatory damages, attorneys’ fees, and civil penalties, for violations involving high-risk AI systems.
The available voting history suggests the bill had substantial support in the Senate, passing final passage 51-8 after also clearing the Rules Committee 18-2. The text and caption indicate a strong policy consensus around preventing algorithmic discrimination and increasing transparency and accountability for AI systems. At the same time, the structure of the bill reflects an effort to balance regulation with innovation by limiting coverage to high-risk systems, excluding many ordinary software tools, and including a developer safe harbor.
The main points of contention are likely the breadth of the compliance regime and the potential burden on AI developers and deployers. The bill imposes mandatory audits, reporting, notice, human-review rights, and private litigation exposure, which could be viewed as costly or difficult to implement, especially for startups and smaller firms. Another likely issue is the scope of “high-risk AI system” and “algorithmic discrimination,” as well as the bill’s strong enforcement provisions, including a private right of action and a presumption against dismissal unless rebutted by clear and convincing evidence. Supporters, by contrast, appear focused on civil rights protections, transparency, and preventing discriminatory automated decisions.