No Robot Bosses Act
- Sponsor
- Rep. Bonamici, Suzanne [D-OR-1]
- Committees
- Oversight and Government Reform Committee (primary)
- Last action
- Dec 3, 2025
Bottom line
The "No Robot Bosses Act" seeks to ensure human oversight and accountability in the use of artificial intelligence and automated systems in employment decisions, aiming to protect workers from algorithmic bias and opaque management practices.
What it actually does
This bill prohibits employers from exclusively relying on automated decision systems (ADS) for employment-related decisions and imposes strict requirements for any use of ADS outputs, including pre-deployment testing, annual independent audits for bias, comprehensive disclosures to covered individuals, human corroboration, and a human-led dispute and appeal process. It also establishes a new Technology and Worker Protection Division within the Department of Labor and invalidates predispute arbitration agreements and joint-action waivers for violations.
Proponents argue
Supporters argue that the rapid adoption of AI in hiring, management, and termination processes poses significant risks of discrimination, unfair treatment, and lack of transparency for workers. They contend that this bill is essential to safeguard fundamental worker rights, ensure human accountability, and prevent the perpetuation or amplification of existing biases through automated systems, thereby promoting fairness and equity in the workplace.
Opponents contend
Critics might argue that the bill imposes overly burdensome compliance requirements on employers, stifling innovation and increasing operational costs, particularly for small and medium-sized businesses. They may contend that strict regulations on AI use could hinder efficiency gains and the adoption of beneficial technologies, potentially making U.S. businesses less competitive globally.
The bill is moderately detailed, requiring careful reading to understand the specific compliance and enforcement mechanisms, but its length is manageable for an informed reader.
Section 3(a)(1)
Prohibiting Exclusive Reliance on Automated Decision Systems
This provision forbids employers from making employment-related decisions solely based on an automated decision system. If an employer uses an ADS output, they must ensure the system has undergone pre-deployment testing for efficacy and compliance with anti-discrimination laws, is independently tested annually for bias, and its output is corroborated by a human.
Supporters argue
Supporters argue this is a critical safeguard against unchecked algorithmic decision-making, ensuring that human judgment and accountability remain central to employment practices and mitigating the risk of systemic discrimination.
Critics contend
Opponents may argue that mandating human corroboration for every ADS output could negate efficiency gains, increase labor costs, and slow down decision-making processes, particularly for large organizations.
Tradeoffs
Balances the desire for efficiency and scale offered by AI with the need for fairness, human oversight, and protection against algorithmic bias in critical employment decisions.
Section 3(a)(2)
Mandating Transparency for ADS Use
Employers must disclose to covered individuals that an ADS is being used, describe the system (data inputs, characteristics measured, how it relates to job functions, how to interpret outputs), identify the system operator, explain how the output is used, and detail the dispute/appeal process. Initial disclosures are required before hiring decisions or application acceptance, with updates for significant changes.
Section 3(a)(3)
Training for Individuals Using Automated Decision Systems
Employers must train individuals who operate or use ADS outputs on the system's input information, appeal processes, potential biases, limitations, adverse effects, errors, and inappropriate uses.
Section 3(b)
Human Manager Option for Automated Management
An employer managing a covered individual through an automated decision system must allow that individual to opt out and instead be managed by a human manager capable of making employment-related decisions.
Section 4
New Division within Department of Labor
Establishes a new Technology and Worker Protection Division within the Department of Labor, headed by an Administrator. The division is authorized to hire technologists and other staff and establish four advisory boards (User, Research, Product, Labor) composed of diverse experts to advise on the Administrator's functions and emerging data practices.
Section 6
Protections Against Retaliation
Prohibits employers from discriminating or retaliating against any covered individual for exercising their rights under this Act, filing complaints, seeking assistance, or participating in any inquiry or proceeding related to the Act.
Section 7(a)
Enforcement by DOL and Private Litigation
Grants the Secretary of Labor (through the new Division) authority to investigate and resolve complaints. It also establishes a private right of action, allowing adversely affected covered individuals or labor organizations to sue employers for violations of Sections 3 or 6, with potential relief including actual damages, treble damages, statutory damages (up to $100,000 for willful retaliation), injunctive relief, and attorney's fees. States and Tribal governments waive sovereign immunity for suits under this section if they receive federal financial assistance.
Section 7(f)
Prohibiting Forced Arbitration for ADS Disputes
This provision declares that no predispute arbitration agreement or predispute joint-action waiver shall be valid or enforceable with respect to any alleged violation of Sections 3 or 6 of the Act.
Defines "automated decision system" broadly to include any system using computation to determine outcomes, make decisions, inform policy, or collect data, derived from machine learning, statistics, or AI. Crucially, it *excludes* "passive computing infrastructure" like web hosting, networking, or data storage that doesn't influence decisions.
Section 2(1)
Why it matters:Standard legislative practice to define terms upfront, but the distinction between "system" and "infrastructure" is critical for scope.
Case for: Supporters argue that a clear and broad definition is necessary to capture the full range of AI tools used in employment, while the exclusion of passive infrastructure prevents overreach and avoids regulating basic IT services.
Case against: Critics might argue that the definition is still too broad and could inadvertently capture systems not intended for regulation, or that the distinction between "active" and "passive" systems could be a source of future litigation.
Estimated impact: Determines the entire scope of the bill's applicability to employer technologies.
Requires automated decision systems to comply with the Artificial Intelligence Risk Management Framework released by the National Institute of Standards and Technology (NIST) on January 26, 2023, or any successor framework.
Section 3(a)(1)(B)(i)(IV)
Why it matters:Technical compliance detail within a list of other compliance requirements.
Case for: Proponents argue that leveraging an existing, widely recognized framework from a credible federal agency provides a standardized, best-practice approach to AI risk management, ensuring systems are developed and deployed responsibly.
Case against: Some might argue that mandating compliance with a specific framework, especially one that can evolve, could create rigidity or administrative burdens, particularly for smaller employers who may lack the resources to fully implement complex frameworks.
Estimated impact: Establishes a technical standard for AI system development and deployment in employment.
States and Tribal governments that receive or use Federal financial assistance for any program or activity waive sovereign immunity to suits brought by covered individuals or labor organizations under this Act.
Section 7(3)(D) and (E)
Why it matters:Technical legal language related to governmental liability, often placed in enforcement details.
Case for: Supporters argue this is essential to ensure that public sector employees have the same protections and recourse as private sector employees, preventing governments from using sovereign immunity to avoid accountability for algorithmic discrimination.
Case against: State and Tribal governments might argue this infringes on their sovereign rights and could expose them to significant litigation risks, potentially making them hesitant to accept federal funding or adopt AI technologies.
Estimated impact: Significantly expands the scope of potential litigation against public sector employers.