Restricts the use of automated decision systems for employment-related decisions, including by requiring such systems to be tested for efficacy, non-discrimination, and relevant laws & standards.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding legislative bill with mandatory language throughout, explicit enforcement mechanisms including civil penalties and private rights of action, and establishment of a federal enforcement agency (Technology and Worker Protection Division).
The document has good coverage of approximately 6-8 subdomains, with strong focus on unfair discrimination (1.1, 1.3), privacy compromise (2.1), overreliance and unsafe use (5.1), loss of agency (5.2), increased inequality (6.2), and lack of robustness (7.3). Coverage is concentrated in discrimination, human-computer interaction, and AI system safety domains, specifically addressing employment-related AI risks.
This legislation governs AI use across all economic sectors where employers use automated decision systems for employment-related decisions. The bill applies universally to employers with 11 or more employees across private sector, public sector, and government entities, making it cross-sectoral in nature with particular emphasis on employment practices rather than sector-specific applications.
The document primarily focuses on the Deploy and Operate and Monitor stages, with significant coverage of Verify and Validate. It addresses pre-deployment testing requirements and ongoing monitoring obligations for automated decision systems used in employment contexts.
The document explicitly defines and covers 'automated decision systems' broadly, encompassing AI models and systems used for employment decisions. It does not specifically mention frontier AI, general purpose AI, foundation models, generative AI, predictive AI, open-weight models, or compute thresholds, as its focus is on employment-related automated decision systems regardless of their technical architecture.
United States Congress
The document is titled 'A Bill' which indicates it is proposed legislation by the United States Congress. The authority is listed as 'United States Congress' in the document information.
Secretary of Labor (through Technology and Worker Protection Division Administrator), Equal Employment Opportunity Commission, State attorneys general, State privacy regulators, Comptroller General (for GAO), Librarian of Congress, Board of Directors of Office of Compliance (for Congressional employees), Director of Office of Personnel Management (for federal employees)
The bill establishes multiple enforcement authorities at federal and state levels, with primary enforcement through the newly created Technology and Worker Protection Division within the Department of Labor, supplemented by state-level enforcement and private rights of action.
Technology and Worker Protection Division (including User Advisory Board, Research Advisory Board, Product Advisory Board, Labor Advisory Board), Secretary of Labor, State attorneys general, State privacy regulators
The bill establishes comprehensive monitoring through the Technology and Worker Protection Division with multiple advisory boards, investigation authority, and annual reporting requirements. State-level monitoring is also enabled through state attorneys general and privacy regulators.
Employers (including private sector employers with 11+ employees, federal agencies, state and local government entities, Congressional offices, Library of Congress, Government Accountability Office, tribal governments)
The bill explicitly targets employers who use automated decision systems for employment-related decisions. The definition of 'employer' is comprehensive and includes various government and private sector entities that employ or engage workers.
8 subdomains (6 Good, 2 Minimal)