AI Policy Wiki
Dashboard

OSTP Blueprint for an AI Bill of Rights (October 2022)

medium confidence · updated 2026-05-11

Biden-administration OSTP white paper outlining five principles for AI use in automated systems affecting Americans — safe and effective systems, algorithmic discrimination protections, data privacy, notice and explanation, human alternatives. Non-binding; supersession trajectory after EO 14365 / America's AI Action Plan is partial and contested.

White House Office of Science and Technology Policy white paper, published October 4, 2022. Outlines five principles for the design, use, and deployment of automated systems that affect Americans' rights, opportunities, or access. Non-binding; provides framework + recommendations rather than enforceable rules.

Why this page exists

Referenced 3+ times across the wiki as the anchoring instrument for the Biden-era civil-rights framing of AI policy. Pairs with Executive Order 14110 (RESCINDED) (rescinded) and the NIST AI RMF as the Biden-administration AI-policy triptych.

The five principles

  1. Safe and effective systems. Systems should be safe and effective at their intended purpose; tested before deployment, ongoing monitoring, retired if unsafe.
  2. Algorithmic discrimination protections. Systems should not contribute to unjustified disparate-impact discrimination; affected groups should be able to seek redress.
  3. Data privacy. Default privacy protections; designs that minimize data collection; informed consent; restrictions on surveillance.
  4. Notice and explanation. People should know when an automated system is being used and how it affects them; understandable explanations.
  5. Human alternatives, consideration, and fallback. Right to opt out where appropriate; human consideration and fallback in critical domains.

Status (2026-05-11)

  • Formally still on the OSTP website, but its operational influence is contested:
    • Superseded in framing by America's AI Action Plan (July 2025) and Executive Order 14365 (December 2025), which take a deregulatory posture.
    • NIST AI RMF revision (America's AI Action Plan Pillar I directs NIST to remove DEI, misinformation, and climate references) effectively narrows the regulatory inheritance.
    • State-level adoption continues — many state AI bills (CA SB 53, Colorado AI Act, NY RAISE Act) inherit framing from the Blueprint.

Influence

  • Reference instrument for state AI legislation. Multiple state-level AI bills cite the Blueprint's five principles, particularly the algorithmic-discrimination and notice-and-explanation principles.
  • Reference instrument for civil-society advocacy. ACLU, EFF, Access Now, and similar organizations frame ongoing AI civil-rights advocacy around Blueprint principles.
  • Reference instrument for court submissions. Some AI civil-rights filings cite the Blueprint as a statement of executive-branch policy at filing time, even if subsequently revised.

Relationships

Sources

This source page is a stub created during the v4.0 backlog-close cycle (2026-05-11). The primary OSTP Blueprint text is queued for foundational ingest as a primary-text source. Save to Raw Sources/ostp-blueprint-ai-bill-of-rights-2022.md and complete the ingest workflow.