Summary
Regulates automated decision-making systems used by New York government agencies through meaningful human review, periodic impact assessments, bias/privacy/cybersecurity/public-health risk review, and public disclosure of agency ADM tools. S07599 created Article 5 of State Technology Law and supersedes the broader S07543 Article 4 structure, which is now narrowed to employment decisions; Article 5 sunsets July 1, 2028.
Healthcare Implications
Directly affects state health, Medicaid, mental-health, and social-service programs that use AI or automated decision systems for eligibility, fraud detection, case prioritization, enforcement, or access to services. Agencies must evaluate public-health/safety risks, bias, privacy, and cybersecurity impacts and maintain meaningful human review for decisions affecting benefits, rights, and welfare.
Operational Implications
- State agencies using ADM/AI in functions affecting public benefits, rights, or welfare must subject the system to continued and operational meaningful human review.
- Periodic impact assessments required every 2 years and before material changes; assessments cover objectives, design/training data, bias, cybersecurity, public-health/safety risks, and data usage.
- If assessment finds discriminatory or biased outcomes, the agency must cease using the system.
- Agencies must publish annual lists of ADM tools and submit impact assessments to state leadership; Article 5 sunsets July 1, 2028.