Algorithmic Accountability Act of 2025
- Bill Number
- H.R. 5511
- Origin Chamber
- House
- Congress
- 119th Congress, Session 1
- Policy Area
- Commerce
- Status
- Introduced
- Latest Action
- 2025-09-19: Referred to the House Committee on Energy and Commerce.
- Last Updated
- 2026-06-26T08:07:16Z
AI-Generated Summary
Purpose
The Algorithmic Accountability Act of 2025 aims to promote accountability in the use of advanced algorithms (like those powered by artificial intelligence or machine learning) that influence important decisions affecting people's lives. It requires large companies to assess and report on the potential harms of these algorithms to consumers, helping to identify and reduce biases, privacy risks, and other negative impacts in areas like employment, healthcare, and housing.
Key Provisions
- Definitions: Establishes terms such as "covered algorithm" (complex computational processes using AI techniques that influence decisions), "covered entity" (generally large businesses with over $50 million in annual revenue or handling data on over 1 million consumers, plus some smaller firms using high-risk algorithms), "critical decision" (choices affecting access to essentials like jobs, education, healthcare, or financial services), and "impact assessment" (ongoing evaluation of an algorithm's effects on consumers).
- Prohibited Acts and Regulations: Makes it unlawful for covered entities to deploy high-risk algorithms without performing impact assessments. The Federal Trade Commission (FTC) must issue rules within 2 years, in consultation with experts, requiring assessments before and after deployment, documentation retention for years after use, annual summary reports to the FTC, stakeholder consultations, and efforts to mitigate harms. Rules also cover disclosure of covered entity status and guidelines for prioritizing assessments.
- Impact Assessment Requirements: Covered entities must evaluate algorithms for privacy risks, performance (including biases based on race, gender, etc.), data sources, consumer rights (like opt-out options and appeals), and negative impacts. They document consultations, testing, training for staff, and any unfeasible requirements with reasons (e.g., lack of data).
- Summary Reports: Entities submit concise reports to the FTC annually (or initially for new algorithms) detailing the algorithm's purpose, performance metrics, consultations, data use, transparency measures, identified harms and mitigations, and infeasible assessments. Reports follow a standardized, machine-readable format.
- Reporting and Public Repository: The FTC publishes annual aggregated reports on trends from submissions and creates a public online repository (updated quarterly) with basic info on algorithms (e.g., company name, decision type, opt-out links) to inform consumers and researchers, while protecting business secrets.
- Guidance and Resources: The FTC provides templates, training, and assistance for compliance, including help determining covered entity status. It establishes a new Bureau of Technology within the FTC (with at least 50 staff) for tech expertise and adds 25 enforcement personnel. Authorizes funding for these.
- Enforcement: Violations are treated as unfair or deceptive practices under existing FTC law, allowing FTC investigations, fines, and remedies. States can sue on behalf of residents (with notice to FTC), and the FTC coordinates with other agencies. No preemption of state or local laws.
Significant Changes to Existing Law
- Introduces mandatory pre- and post-deployment impact assessments for algorithms in critical decisions, which were previously unregulated at the federal level beyond general consumer protection laws.
- Expands FTC authority to regulate AI-driven decisions, treating non-compliance as a violation of the Federal Trade Commission Act (15 U.S.C. § 41 et seq.), enabling broader enforcement tools like civil penalties.
- Creates new reporting obligations and a public repository, shifting from voluntary industry practices to required transparency without mandating full disclosure of proprietary details.
- Adjusts thresholds for applicability (e.g., revenue and data-handling limits, with inflation adjustments) to target larger or high-impact users, differing from sector-specific rules (e.g., in finance or healthcare).
Potential Impacts
- Government Agencies: Increases FTC workload with rulemaking, report reviews, and a new bureau, requiring additional staff and funding; promotes coordination with agencies like NIST (National Institute of Standards and Technology) for standards. States gain enforcement tools, potentially leading to more litigation.
- Citizens: Enhances protections against discriminatory or harmful algorithmic decisions in daily life areas, improving transparency (e.g., appeal rights) and privacy; the public repository aids awareness and recourse, though benefits depend on FTC implementation.
- International Relations: Minimal direct impact, but could influence global AI standards by setting a U.S. model for accountability, potentially affecting trade with countries having stricter (e.g., EU AI Act) or laxer regulations; no explicit international provisions.
- Businesses: Imposes compliance costs (assessments, reporting) on covered entities, possibly slowing innovation or raising barriers for smaller firms, but encourages better practices like bias mitigation; voluntary submissions allowed for non-covered entities.
Main Stakeholders Affected
- Covered Entities: Primarily large tech, finance, healthcare, and other companies deploying AI algorithms for critical decisions (e.g., hiring tools, loan approvals); they face assessment and reporting duties.
- Consumers: Individuals impacted by algorithmic decisions in employment, housing, etc., gaining protections against biases and more recourse options.
- FTC and Regulators: FTC leads implementation and enforcement; states' attorneys general can act; other agencies (e.g., NIST, OSTP) provide input and access to reports.
- Advocates and Researchers: Civil rights groups, academics, and tech experts benefit from consultations, public data, and trends reports to study and improve AI fairness.
- Employees and Internal Teams: Company staff involved in algorithm development must receive training and document processes.
Notable Legal, Constitutional, or Political Implications
- Legal: Strengthens consumer protection under FTC Act by addressing emerging AI risks without creating a new agency; allows state enforcement to foster a patchwork of regulations, potentially increasing compliance complexity but avoiding federal overreach. Emphasizes mitigation of "material negative impacts" while permitting unmitigated harms if justified by non-discriminatory interests.
- Constitutional: Supports equal protection by requiring bias evaluations (e.g., on race, gender), aligning with anti-discrimination principles; balances free speech/business rights by limiting disclosures to summaries, avoiding compelled proprietary revelations.
- Political: Represents bipartisan potential in AI regulation (introduced by Democrats but focuses on accountability over bans); could spark debates on innovation stifling vs. harm prevention, especially amid growing AI concerns; periodic reviews ensure adaptability, but funding reliance on Congress may limit effectiveness.
This summary was generated by AI and may contain inaccuracies. Refer to the official source document for the authoritative text.
Sponsor
Rep. Clarke, Yvette D. [D-NY-9]
Cosponsors (30)
Rep. Balint, Becca [D-VT-At Large], Rep. Barragán, Nanette Diaz [D-CA-44], Rep. Bell, Wesley [D-MO-1], Rep. Brown, Shontel M. [D-OH-11], Rep. Davis, Danny K. [D-IL-7], Rep. Deluzio, Christopher R. [D-PA-17], Rep. Evans, Dwight [D-PA-3], Rep. Figures, Shomari [D-AL-2], Rep. Foushee, Valerie P. [D-NC-4], Rep. García, Jesús G. "Chuy" [D-IL-4], Del. Norton, Eleanor Holmes [D-DC-At Large], Rep. Huffman, Jared [D-CA-2], Rep. Jackson, Jonathan L. [D-IL-1], Rep. Jacobs, Sara [D-CA-51], Rep. Jayapal, Pramila [D-WA-7], Rep. Kelly, Robin L. [D-IL-2], Rep. Lee, Summer L. [D-PA-12], Rep. Ramirez, Delia C. [D-IL-3], Rep. Tlaib, Rashida [D-MI-12], Rep. Veasey, Marc A. [D-TX-33], Rep. Wilson, Frederica S. [D-FL-24], Rep. Espaillat, Adriano [D-NY-13], Rep. Cohen, Steve [D-TN-9], Rep. Trahan, Lori [D-MA-3], Rep. Pressley, Ayanna [D-MA-7], Rep. Salinas, Andrea [D-OR-6], Rep. Ocasio-Cortez, Alexandria [D-NY-14], Rep. Omar, Ilhan [D-MN-5], Rep. McClellan, Jennifer L. [D-VA-4], Rep. Hayes, Jahana [D-CT-5]
Recent Actions
- 2025-09-19: Referred to the House Committee on Energy and Commerce.
- 2025-09-19: Introduced in House
- 2025-09-19: Introduced in House
Bill Versions
- Algorithmic Accountability Act of 2025 — issued 2025-09-19 — PDF (47 pages)