Policy Archive · Episode 03

The Case of the Government Goof-Up

Explore the real-world policies behind this case. Follow proposed and existing policy, potential gaps, and outcomes as they develop.

• Proposed Policy Proposed — Not Enacted

  H.R. 9125 — Sectoral AI Governance Act of 2026

  H.R. 9125 would authorize federal agencies that enforce federal law to issue rules regulating algorithmic decision-making systems when those systems are likely to materially contribute to violations of laws the agency enforces.

  Identifier

  H.R. 9125

  Jurisdiction

  United States (Federal)

  Status

  Proposed — Not Enacted

  Dates

  H.R. 9125 was introduced in the House on June 3, 2026 and referred to the House Committees on the Judiciary and Oversight and Government Reform. It has not been enacted.

  Latest update

  June 3, 2026

  Update date

  2026-06-03

  What it does: H.R. 9125 would authorize federal agencies that enforce federal law to issue rules regulating algorithmic decision-making systems when those systems are likely to materially contribute to violations of laws the agency enforces. The bill also requires agencies to consider potential disruption to government services or public benefits when developing those rules.

  Potential impact: The proposal could give federal agencies clearer authority to regulate uses of algorithmic decision-making that may contribute to violations of federal law. It would not itself give individuals specific rights to notice that AI was used, explanation of an AI-supported decision, correction, human review, appeal, or redress.

  Sources

  • Congress.gov — H.R. 9125, Sectoral AI Governance Act of 2026
  • GovInfo — H.R. 9125 (Introduced in House)
• Existing Policy In Effect

  Federal Agency AI Governance — AI in Government Act and OMB M-25-21

  The AI in Government Act of 2020 established a federal framework for agency use of AI, and OMB Memorandum M-25-21 provides current operational guidance for federal agency use of AI, including safeguards for high-impact AI.

  Identifier

  OMB M-25-21

  Jurisdiction

  United States (Federal)

  Status

  In Effect

  Dates

  The AI in Government Act was enacted in 2020. OMB M-25-21 was issued on April 3, 2025 and remains current federal policy governing Executive Branch agency use of AI.

  Latest update

  AI in Government Act — 2020 OMB M-25-21 — April 3, 2025

  Update date

  2025-04-03

  What it does: The AI in Government Act of 2020 established a federal framework for agency use of AI, including government-wide guidance and agency planning. OMB Memorandum M-25-21, issued April 3, 2025, provides current operational guidance for federal agency use of AI. For AI classified as high-impact, M-25-21 requires substantial agency-side safeguards, including pre-deployment testing, AI impact assessments and ongoing monitoring. It also addresses governance, public AI inventories, risk mitigation and discontinuation when risks cannot be adequately mitigated. The memorandum says affected individuals should have access to timely human review and remediation, but it does not establish an enforceable individual right to that process.

  Potential impact: Together, these policies provide an existing framework for managing AI inside federal agencies. Their safeguards primarily govern what agencies must do to manage AI risk. They do not provide a general set of individual, AI-specific rights to notice, explanation, correction, human review, or redress when an AI-supported government decision affects a person. General protections under administrative law, constitutional due process, the Privacy Act, and program-specific appeals may still apply depending on the circumstances.

  Sources

  • GovInfo — Consolidated Appropriations Act, 2021 (includes the AI in Government Act of 2020)
  • Office of Management and Budget — M-25-21, Accelerating Federal Use of AI through Innovation, Governance, and Public Trust — OMB
• Policy Gap Partial Gap

  Government AI Error, Accountability and Redress

  Jurisdiction

  United States (Federal)

  Status

  Partial Gap

  What it does: Existing law provides general ways to challenge some government decisions or correct government records, and M-25-21 provides substantial agency-side safeguards for high-impact AI. But Judge GPT TV's research did not identify a general federal framework giving individuals AI-specific rights to be notified that AI contributed to a decision about them, receive an explanation of its role, correct an AI-related error, obtain enforceable human review, or receive defined redress.

  Potential Policy Question

  When government AI contributes to a consequential decision about a person, what rights should that person have to understand, challenge, and correct the decision?

  Sources

  • Office of Management and Budget — M-25-21 — OMB
  • U.S. House Office of the Law Revision Counsel — Privacy Act, 5 U.S.C. § 552a — OLRC
• Policy Gap Partial Gap

  Safeguards for High-Impact Government AI

  Jurisdiction

  United States (Federal)

  Status

  Partial Gap

  What it does: M-25-21 already requires significant internal safeguards for high-impact federal AI, including testing, impact assessment, monitoring, risk mitigation, governance and documentation. However, these safeguards are primarily administered within the Executive Branch. The current framework does not provide a general system of individual-facing notice and explanation or an independent external mechanism through which affected individuals can enforce M-25-21's AI safeguards.

  Potential Policy Question

  What additional accountability should apply when federal agencies use high-impact AI in decisions that can significantly affect individuals?

  Sources

  • Office of Management and Budget — M-25-21 — OMB
  • GovInfo — AI in Government Act of 2020