Require the state to conduct ADS Impact Assessments before use. Mandate neutral third-party evaluations on impacts. Prohibit ADS purchase/use without publishing the assessment and public comments. Obligate a list of qualified assessors. Demand plain-language notifications and human review mechanisms. Ban payments for non-compliant ADS.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding legislative act with mandatory requirements, enforcement mechanisms, penalties, and legal remedies. The document uses mandatory language throughout ('shall', 'must', 'required') and establishes specific legal obligations for state agencies with enforcement through payment prohibitions and private rights of action.
The document has good coverage of approximately 10-12 subdomains, with strong focus on discrimination and bias (1.1, 1.3), privacy compromise (2.1), security vulnerabilities (2.2), malicious use for surveillance (4.1), overreliance (5.1), loss of agency (5.2), governance failure (6.5), lack of robustness (7.3), and lack of transparency (7.4). Coverage is concentrated in fairness, transparency, accountability, and human oversight domains.
This document governs automated decision system use exclusively within the Public Administration sector. It applies to all New York State government agencies, political subdivisions, and public benefit corporations across all their functions. The document does not regulate private sector use of AI in any industry.
The document covers multiple lifecycle stages with primary focus on Verify and Validate, Deploy, and Operate and Monitor stages. It mandates impact assessments before deployment, requires ongoing monitoring and audits, and establishes human review mechanisms for operational systems. The Plan and Design stage is implicitly covered through requirements for system documentation and design considerations in impact assessments.
The document focuses on 'Automated Decision Systems' broadly defined as software designed to aid or replace human decision-making. It does not explicitly mention AI models, AI systems, frontier AI, general purpose AI, foundation models, generative AI, or compute thresholds. The scope is functional rather than technical, focusing on systems that impact human welfare and rights regardless of underlying technology.
New York State Legislature
This is a legislative act ('An Act to amend the general business law, the executive law, the state finance law and the education law') proposed by the New York State Legislature to establish the Digital Fairness Act.
Committee on Open Government, State Procurement Council, Courts of competent jurisdiction, City or County Council, State Legislature, Office of Information Technology Services, Division of Human Rights
Multiple government bodies are designated with enforcement authority including conducting audits, maintaining qualified assessor lists, approving system acquisitions, and adjudicating violations. Courts have jurisdiction to award damages and injunctive relief.
Committee on Open Government, Office of Information Technology Services, Division of Human Rights, neutral third-party assessors, the public (through comment periods), affected communities
The document establishes comprehensive monitoring through annual audits, third-party impact assessments, public comment periods, and ongoing review by multiple government agencies. The public and affected communities have formal roles in monitoring through comment periods and community engagement.
The State of New York and any governmental agency, political subdivision or public benefit corporation of the State; vendors of automated decision systems
The document explicitly targets state government entities that purchase, deploy, or use automated decision systems, as well as vendors who provide such systems. The requirements apply to government agencies as deployers of AI systems.
12 subdomains (8 Good, 4 Minimal)