Government AI Actors Overview

Quick Assessment

Key Links

Overview

Government AI actors are the organizations and officials who shape how artificial intelligence is developed, deployed, governed, and constrained within and by public institutions. They span every branch and level of government — from heads of state issuing executive orders, to parliaments passing legislation, to technical standards bodies writing interoperability specifications, to courts adjudicating liability when AI systems cause harm. Collectively, they constitute the public counterpart to the private sector ecosystem of labs, investors, and deployers that currently leads much of frontier AI development. For a broader mapping of who holds power over AI trajectories, see the US Government Authority Over Commercial AI InfrastructureAnalysisUS Government Authority Over Commercial AI InfrastructureSurveys US legal authority (DPA, IEEPA, CLOUD Act, FISA 702) over $700B+ in commercial AI infrastructure concentrated in 5-6 companies, concluding the government has extensive but not unlimited pow...Quality: 64/100 analysis and the AI Power and Influence Map.

The scale of government AI activity has expanded dramatically since 2023. U.S. federal agencies reported over 1,700 AI use cases in 2024 — more than double the 2023 count — and at least 57 agencies had appointed Chief AI Officers by mid-2024.¹ Globally, AI mentions in legislation rose 21.3% across 75 countries in 2024, representing a ninefold increase since 2016.² Despite this activity, many analysts argue that industry labs still make the most consequential decisions about frontier AI training and deployment, with governments playing a secondary — though increasingly assertive — role.³ The Government Regulation vs Industry Self-GovernanceCruxGovernment Regulation vs Industry Self-GovernanceComprehensive comparison of government regulation versus industry self-governance for AI, documenting that US federal AI regulations doubled to 59 in 2024 while industry lobbying surged 141% to 648...Quality: 54/100 debate remains central to evaluating the effectiveness of these actors.

This article organizes government AI actors by function: executive bodies that set strategic direction; legislatures that authorize and constrain; regulatory agencies that implement and enforce; standards bodies that establish technical norms; and judicial actors that adjudicate disputes. Each category operates through distinct power mechanisms and at different speeds, creating a layered and sometimes inconsistent governance architecture.

History

Government engagement with AI as a distinct policy domain began in earnest around 2016–2018, when several national AI strategies were published and bodies like the U.S. Select Committee on Artificial Intelligence were established. However, the governance infrastructure remained sparse and largely advisory until the early 2020s, when large language models and generative AI moved from research settings into mass deployment.

The United Kingdom announced a 10-year plan to become a global AI superpower in 2021, emphasizing R&D investment and industry adoption.⁴ That same year, formal U.S. policy development discussions began at the federal level. A significant acceleration occurred in 2023, when President Biden issued an executive order directing federal agencies to develop guidelines, standards, and best practices for AI safety and security — the broadest U.S. federal AI directive to that point.⁴ In February 2024, the UK's Office for Artificial Intelligence was absorbed into the newly created Department for Science, Innovation and Technology, consolidating oversight.⁴

The U.S. Office of Management and Budget issued Memorandum M-24-10 in 2024, establishing the first government-wide policy framework for federal agency AI use, requiring risk assessments, transparency measures, and Chief AI Officer appointments.⁵ By mid-2025, the Trump administration had released an AI Action Plan and signed three implementing executive orders, revising the NIST AI Risk Management Framework and formalizing an interagency Chief AI Officer Council.¹ Internationally, the EU AI Act entered enforcement for prohibited practices in 2025, and China continued advancing its own governance regime through the Cyberspace Administration of China.

The OECD's 2025 Digital Government Index found that 70% of countries had used AI to improve internal governmental processes, but only 33% had deployed AI in citizen-facing services — a gap reflecting uneven readiness and risk tolerance across governments.⁶

Key Activities

1. Executive Bodies: Strategic Direction and Mandate

Executive actors — heads of state, cabinets, and their central policy offices — set the overall direction for national AI strategy, issue binding directives to agencies, and signal geopolitical priorities through investment commitments.

United States: The White House has been the most active executive AI actor in the democratic world. Biden's 2023 executive order directed agency participation in AI safety guidelines; his OMB subsequently issued M-24-10 and M-25-21. The Trump administration's 2025 AI Action Plan shifted emphasis toward innovation, open-source models, and deregulation, including revisions to NIST frameworks that removed references to misinformation, DEI, and climate change.¹ The Department of Government Efficiency (DOGE), established under a Trump executive order, placed AI at the center of federal technology modernization.² The administration also announced the Stargate Project, a $500 billion joint venture with SoftBank, OpenAIOrganizationOpenAIComprehensive organizational profile of OpenAI documenting evolution from 2015 non-profit to Public Benefit Corporation, with detailed analysis of governance crisis, 2024-2025 ownership restructuri...Quality: 62/100, Oracle, and MGX to build AI infrastructure.²

European Union: The European Commission developed the EU AI Act — the world's first comprehensive AI law — which began enforcement of prohibited practices in 2025. The Commission's Directorate-General for Communications Networks, Content and Technology (DG CNECT) leads implementation. The Commission also produced the Assessment List for Trustworthy Artificial Intelligence (ALTAI) framework.³

United Kingdom: The Cabinet Office and Department for Science, Innovation and Technology jointly oversee UK AI strategy. The UK hosted the 2023 AI Safety Summit at Bletchley Park, producing an international declaration, and has positioned itself as a facilitator of global AI governance dialogue.

China: The Chinese Communist Party Politburo and State Council direct AI policy through bodies including the Ministry of Science and Technology and the Cyberspace Administration of China. China committed a $47.5 billion semiconductor fund as part of its AI infrastructure strategy.² DeepSeek-R1's January 2025 release demonstrated that Chinese actors could advance frontier AI capabilities while circumventing U.S. chip export controls through efficiency gains.²

Other significant national investments include France (€109 billion), Canada ($2.4 billion), India ($1.25 billion), and Saudi Arabia ($100 billion Project Transcendence).²

2. Legislative Bodies: Authorization, Investigation, and Constraint

Legislatures authorize AI programs, appropriate funding, hold hearings, and pass laws that constrain both government and private AI use. They typically act more slowly than executive actors but produce durable legal frameworks.

U.S. Congress: Congress has been active in oversight hearings but slower to pass comprehensive AI legislation. By 2024, federal agencies had issued 59 AI-related regulations — double the 2023 count, from twice as many agencies.² The TAKE IT DOWN Act, signed in May 2025, became the first federal law targeting AI-generated non-consensual intimate imagery, passing the House 409–2.⁷ Congress also exercises indirect influence through appropriations and confirmation authority over agency leadership.

European Parliament: The Parliament was a co-legislator on the EU AI Act, pushing for stronger prohibitions and rights protections during negotiations with the Council. It continues to shape implementation through committee oversight and secondary legislation.

UK Parliament: Westminster oversight occurs primarily through select committees (including the Science and Technology Committee) and Lords inquiries. The UK has not passed comprehensive AI primary legislation, instead relying on sector-specific regulators under existing legal frameworks.

China's National People's Congress (NPC): The NPC enacts enabling legislation for AI governance, with the Cyberspace Administration of China handling detailed regulatory implementation. The NPC's role is more ratificatory than deliberative, reflecting the CCP's dominant executive position.

U.S. State Legislatures: State-level activity has been substantial. California signed 18 AI bills in 2024 affecting consumer products, healthcare, government, and employment. New York's Local Law 144 addresses AI in hiring decisions. Colorado passed the Colorado AI Act establishing developer obligations. The National Conference of State Legislatures tracks AI legislation across healthcare, elections, workforce, and criminal justice domains.⁸

3. Regulatory Agencies: Implementation and Enforcement

Regulatory agencies translate legislative mandates and executive directives into operational rules. They conduct inspections, issue guidance, impose penalties, and build sector-specific AI governance capacity.

NIST (National Institute of Standards and Technology): NIST published the AI Risk Management Framework (AI RMF), which has become the most widely referenced U.S. government standard for AI governance. Its appendix on AI Actor Tasks defines governance roles across design, deployment, impact assessment, and oversight functions.⁹ The 2025 executive orders directed revisions to the RMF, raising concerns among some practitioners about the removal of references to bias and misinformation risks.¹

Department of Commerce / Bureau of Industry and Security (BIS): BIS administers export controls on advanced semiconductors and AI-related technology, a central tool in the U.S.-China AI competition. These controls have had significant effects on AI hardware supply chains, though China's development of efficient models like DeepSeek-R1 has complicated the strategy.²

Cyberspace Administration of China (CAC): The CAC regulates AI content generation, recommendation algorithms, and generative AI services in China. It has issued detailed rules on deepfakes, algorithmic transparency, and training data requirements.

UK Competition and Markets Authority (CMA): The CMA has examined AI foundation models for competition law implications and coordinates with other UK regulators on AI oversight under a principles-based, sector-led approach.

Federal Trade Commission (FTC): The FTC applies existing consumer protection and competition law to AI products, with enforcement actions focused on deceptive AI claims, biometric data misuse, and unfair algorithmic practices.

Office of Management and Budget (OMB): While primarily an executive office, OMB functions as a key regulatory actor by issuing binding guidance to federal agencies. M-24-10 established high-risk AI requirements including mandatory risk assessments, transparency disclosures, and impact safeguards across agencies.⁵

Agency-Level Implementation: Larger agencies such as the Department of Homeland Security and Department of Labor have built AI inventories, use-case documentation, and governance plans. Smaller agencies — the research notes the Court Services and Offender Supervision Agency as an example — struggle with compliance due to resource constraints.⁵

4. Standards Bodies: Technical Norms and Interoperability

Standards bodies produce voluntary technical specifications that shape how AI systems are built and evaluated, often carrying significant de facto regulatory weight.

NIST AI RMF: Though NIST is a federal agency, its standards function is analytically distinct. The AI RMF defines terminology, risk categories, and governance structures used by both public and private actors globally.

ISO/IEC JTC1: The joint technical committee of the International Organization for Standardization and the International Electrotechnical Commission develops international AI standards including ISO/IEC 42001 (AI management systems) and standards on bias, transparency, and safety.

IEEE: The Institute of Electrical and Electronics Engineers develops technical standards for AI systems and has published extensive work on ethically aligned design, though its standards are advisory.

These bodies operate through multi-stakeholder consensus processes that are slower than regulatory timelines but produce technically rigorous outputs with broad international legitimacy.

5. Judicial Actors: Accountability and Rights

Courts adjudicate AI-related disputes and, over time, establish precedents that shape accountability structures. Judicial AI governance is still in early stages but is growing in significance.

Key legal issues include: whether government agencies can disclaim responsibility for harms caused by third-party AI systems they deploy; whether private vendors supplying AI for government functions (welfare eligibility, criminal risk assessment) should be treated as state actors subject to constitutional liability; and whether existing administrative law frameworks adequately constrain algorithmic decision-making in benefits and enforcement contexts.

Some legal scholars have proposed applying state action doctrine — through public function, compulsion, or joint participation tests — to hold AI vendors accountable for constitutional violations when their systems are used in core government functions like welfare administration or pretrial risk assessment.¹⁰ Courts have not yet systematically adopted this framework, but litigation is increasing. The Government Accountability ProjectOrganizationGovernment Accountability ProjectNonprofit organization supporting whistleblower protections. Advocates for AI industry whistleblower rights and transparency in frontier AI development. has tracked cases involving algorithmic accountability in government contexts.

Comparative Overview

Criticisms and Concerns

Several systemic concerns affect government AI actors across all categories. First, accountability gaps are substantial: governments frequently deploy AI systems built by private vendors for high-stakes decisions — benefits eligibility, criminal risk scoring, immigration — while disclaiming responsibility when those systems cause harm, citing limited understanding of the underlying technology.¹⁰ This creates a situation where neither the government deployer nor the private developer faces clear accountability.

Second, public trust is low. Survey data cited in the research suggests 77% of Americans distrust government AI use, 62% lack confidence in federal regulation, and only 44% trust the U.S. government overall on this issue.¹¹ International comparisons show dramatically higher trust in some other national governments, suggesting the deficit reflects specific U.S. institutional conditions rather than a universal public reaction to AI governance.

Third, resource disparities across agencies create uneven governance quality. Large agencies with dedicated AI offices can comply with OMB requirements and build governance infrastructure; smaller agencies cannot, creating systematic blind spots in federal AI oversight.⁵

Fourth, regulatory capture concerns arise at the state level. Governor Newsom's veto of California's AI safety bill was cited as an example of industry influence over state-level governance, while critics of the Trump administration's AI Action Plan argued it subordinated public interest safeguards to the preferences of tech executives who had political proximity to the administration.¹¹

Fifth, the pilot trap — the tendency for government AI initiatives to remain in limited trials without scaling — affects many jurisdictions. Research suggests 70–85% of AI projects fail to meet their stated outcomes, with public sector initiatives at higher risk due to data silos, skills gaps, and procurement rigidities.¹² The Government AI Use MonitoringApproachGovernment AI Use MonitoringSystematic tracking and transparency efforts focused on how governments deploy AI systems, particularly for surveillance, enforcement, and decision-making that affects citizens' rights. Became crit... approach addresses some of these implementation gaps.

Finally, some observers note that government AI governance has focused overwhelmingly on near-term operational and rights-based risks — bias, surveillance, fraud — while giving little attention to longer-term or more speculative risks including AI misalignment and the concentration of AI-enabled power. The Government Regulation vs Industry Self-GovernanceCruxGovernment Regulation vs Industry Self-GovernanceComprehensive comparison of government regulation versus industry self-governance for AI, documenting that US federal AI regulations doubled to 59 in 2024 while industry lobbying surged 141% to 648...Quality: 54/100 debate touches on whether governments are structurally capable of addressing risks at the frontier of AI capability development.

Key Uncertainties

• Whether the U.S. federal governance architecture will remain stable given executive-level policy oscillations between administrations
• How effectively the EU AI Act will be enforced across member states and applied to non-EU actors
• Whether judicial doctrine will evolve to close the vendor accountability gap in government AI deployments
• Whether standards bodies can produce relevant technical norms at the pace of AI capability development
• The extent to which China's governance regime provides genuine public accountability versus serving primarily as a political control mechanism
• Whether government AI maturity — currently at 6% "ideal" state according to one global study — will improve significantly as governance frameworks mature¹³

Sources