Levels of governance

Level 1 of 6

Community and organisational use

Where schools, clinics, community and faith groups, small businesses and households decide whether and how to use AI tools they did not build.

Entities operating here

Over-reliance and deskilling Homogenisation and loss of authorship

What is distinctive here

Little capacity to evaluate a vendor's claims, test a tool on local users, or read a contract's data terms.
Dependence on decisions made elsewhere: the model, its updates and its data practices are set by the vendor, and many rules are set at sectoral or national level.
Close relationships: the people deciding often know the people affected, which can make concerns easier to raise and harder to escalate.
Informal practice governs first: default settings, habits and who is allowed to use a tool shape outcomes long before any written policy exists.

Example. A school can choose an AI tutor's default settings and whether pupils can reach a teacher. It cannot set the vendor's data practices, and it cannot change national curriculum rules. Governance here is real but bounded, and an analysis should show where the boundaries fall.

Scope. Includes households and individuals as a sub-case. A family choosing a child's settings, or a person deciding how far to rely on a chatbot, is governing a use with even less information and leverage than an organisation. Individuals who are affected by others' decisions are covered under the distribution and legitimacy tests, not as a level.

How it meets other levels

Corporate: The vendor's own governance sets what a community user can see and change: model behaviour, updates, data retention and the terms offered.
Sectoral: Sector bodies such as education or health authorities can set rules, guidance and approved-supplier lists that individual organisations rely on because they cannot assess tools themselves.
National: National data-protection, consumer and procurement law sets the floor, and public programmes can fund evaluation or shared capacity that no single organisation could afford.

interpretationConfidence: mediumContestability: medium

Where schools, clinics, community groups and small businesses adopt AI, governance is real but bounded: they can set how a tool is used, but usually cannot evaluate it independently or change how it was built, so their choices depend on decisions made by vendors and higher levels.

Our reading: AIGovSpec's interpretation, added when the community and organisational level was introduced (October 2026). The cited passages concern implementation capacity, dependence on a few providers, and what a well-resourced buyer can demand in a contract; none studies schools or community groups directly. Evidence on governance capacity in small organisations is thin and mostly from high-income settings.

Inspect evidence · 3 passages

Passage summaries are paraphrases, not quotations. Open the source to read the full passage.

Governing AI for Humanity: Final Report

Executive summary, para. xviii (p. 8)

Representation and coordination are said not to be enough: accountability requires implementation so that commitments translate into tangible outcomes, including capacity development, much of it at national and regional levels.

Scope note: An advisory body report representing a majority consensus of its members in their personal capacity. It is not a UN member-state decision. Page locators refer to the PDF's printed page numbers.

Passage checked 2026-09-06 · un-hlab-implementation

M-25-22: Driving Efficient Acquisition of Artificial Intelligence in Government

Section 4(d), ongoing testing and monitoring; Section 4(c)(ii), vendor lock-in

AI contracts must let the buying agency keep monitoring and evaluating the system’s performance, and should include terms such as knowledge transfer that reduce the risk of vendor lock-in.

Scope note: Rescinds and replaces M-24-18. Applies to federal executive agencies; excludes national security systems.

Passage checked 2026-10-09 · omb-m-25-22-monitoring

2023 Landscape: Confronting Tech Power

Executive summary, “Computing Power Advantage” (about p. 7 of the PDF)

Building large AI systems depends on cloud and compute that only a handful of firms run themselves, so other AI companies rely on those incumbents.

Scope note: Civil-society voice. An advocacy report with an explicit position on concentration in the technology sector.

Passage checked 2026-10-09 · ai-now-compute-advantage

Level 2 of 6

Corporate

Allocates responsibility inside firms: boards, safety teams, release decisions, internal escalation.

Entities operating here

AI developer or provider Deploying organization Deployment pressure Competitive race dynamics Internal safety and governance team Evaluation capacity Risk management process Pre-deployment evaluation and red teaming Incident reporting and post-deployment monitoring Over-reliance and deskilling Sycophancy and engagement-driven agreement Homogenisation and loss of authorship Data work and content moderation

Instruments operating at this level

NIST AI RMF 1.0 · voluntary

Level 3 of 6

Sectoral

Addresses domain-specific risks: health, employment, credit, policing, education.

Entities operating here

Deploying organization Procurement leverage Over-reliance and deskilling

No jurisdiction view or instrument is recorded at this level yet. The worked analyses show the level in use.

Level 4 of 6

National

Uses law, procurement, industrial policy, competition policy, and public investment.

Instruments operating at this level

OMB M-25-22 AI acquisition memo · binding
IndiaAI Mission · voluntary
Canada Directive on Automated Decision-Making · binding

Voluntary cross-sector risk-management reference

United States — federal practice layer

NIST's AI RMF supplies a shared practice vocabulary for organizations, but it is not itself a generally binding AI law.

Coverage note: This prototype view is deliberately narrow and does not represent the full federal, state, local, or sector-specific landscape.

National Institute of Standards and Technology

Level 5 of 6

Regional

Harmonizes rules and markets across countries, as the European Union does.

Entities operating here

Liability exposure Regulator or public authority Energy, water, and minerals Supervision and regulatory review Complaint and appeal routes Review and sunset provisions Governance lock-in

Instruments operating at this level

EU Artificial Intelligence Act · binding
EU Product Liability Directive · binding

Supranational market regulation with national and EU-level enforcement

European Union

A binding, risk-based legal architecture that allocates obligations across regulated actors and categories of AI use and models, supervised by designated authorities.

Coverage note: Application is phased and continues to evolve through implementation measures; dates are not hard-coded into this summary.

European Union European Commission

Level 6 of 6

International

Addresses cross-border coordination, shared standards, scientific assessment, security, and representation.

Entities operating here

Civil society and independent researchers Competitive race dynamics Standards body International institution Compute Data, language resources, and data labor Complaint and appeal routes Review and sunset provisions Compute controls Broadly shared benefit Concentration of power Data work and content moderation

Instruments operating at this level

OECD AI Principles · recommendatory
UNESCO AI Ethics Recommendation · recommendatory
Council of Europe Framework Convention on AI · binding
ISO/IEC 42001 AI management system · voluntary

Intergovernmental principles, recommendations, and one treaty

International norms and treaties

OECD and UNESCO instruments create shared vocabularies and normative direction; the Council of Europe Convention adds binding obligations for its parties. Representation in these bodies is uneven.

Coverage note: International alignment does not erase differences in legal authority, institutional capacity, or political legitimacy, and most countries are party to none of the prominent initiatives.

Organisation for Economic Co-operation and Development UNESCO Council of Europe United Nations High-level Advisory Body on Artificial Intelligence