Examples
Worked examples
- Is an instance
A university research computing office adopts an AI-assisted data-analysis tool for a multi-site study. To align with the OECD's robustness/security/safety and accountability principles, it documents who is accountable for validating the tool's outputs, what testing was done before deployment, and what happens if the tool produces an erroneous result in a dataset that informs a publication — rather than treating vendor assurances alone as sufficient.
- Is an instance
A funding agency updates its grant-review guidance to require applicants disclose any generative-AI use in proposal preparation. This operationalizes the transparency and explainability principle at the point where an AI system's involvement could otherwise be invisible to reviewers, consistent with the OECD's 2024-updated emphasis on generative AI.
Counter-examples
Looks similar, but isn't
- Not an instance
A lab publishes a one-line internal memo stating 'AI tools must be used responsibly' with no documented risk assessment, no named accountable owner, and no disclosure mechanism for affected parties. This gestures at the Principles' language without satisfying any of the five values in an operationally checkable way, and would not constitute alignment with the OECD framework as the OECD itself describes it (a lifecycle-wide set of obligations for AI actors, not a values statement).
Editorial commentary
The OECD AI Principles are a set of intergovernmental standards for trustworthy artificial intelligence, adopted by the Organisation for Economic Co-operation and Development’s Council at Ministerial level in May 2019 and updated in May 2024. They were the first AI principles endorsed at the intergovernmental level and were later reflected in the G20 AI Principles. As of the 2024 update, 47 countries and jurisdictions adhere to them, including every OECD member state, the European Union, and a number of non-member adherents (source: OECD, AI principles and OECD.AI, AI Principles overview).
The five values-based principles
- Inclusive growth, sustainable development and well-being — AI should benefit people and the planet.
- Human rights and democratic values, including fairness and privacy — AI actors should respect the rule of law, human rights, and democratic values throughout the AI system lifecycle.
- Transparency and explainability — AI actors should provide meaningful information appropriate to the context, enabling those affected to understand and, where appropriate, challenge outcomes.
- Robustness, security and safety — AI systems should function appropriately and not pose unreasonable safety risk throughout their lifecycle.
- Accountability — AI actors should be accountable for the proper functioning of AI systems, in line with the other four principles.
The five recommendations for policymakers
- Investing in AI research and development.
- Fostering an inclusive AI-enabling ecosystem.
- Shaping an enabling, interoperable governance and policy environment for AI.
- Building human capacity and preparing for labour-market transition.
- International co-operation for trustworthy AI.
The 2024 update
The May 2024 revision responded directly to the emergence of general-purpose and generative AI. It expanded the human-centred-values principle to address AI-amplified misinformation and disinformation while respecting freedom of expression, and sharpened language on privacy, intellectual property, and safety in light of generative-AI-specific risks (source: OECD press release, May 2024).
Why this differs from the EU AI Act
The OECD AI Principles are voluntary, non-binding policy guidance adopted by governments — they set a shared normative baseline and inform national AI strategies, but they create no legal obligations of their own and no enforcement mechanism. The EU AI Act is binding EU law with defined obligations, exemptions (including specific carve-outs relevant to research), and penalties. In practice the two are closely related: the EU AI Act’s own risk-based, human-rights-oriented framing draws on the same normative ground the OECD Principles established in 2019. A research organization operating in the EU needs to track AI Act compliance as a legal matter; the OECD Principles are useful as the broader governance vocabulary and as the operative standard where no binding law yet applies (most jurisdictions outside the EU).
How research organizations apply the Principles
The OECD Principles are not sector-specific, but research institutions, funders, and research-performing organizations increasingly use them as a reference point for institutional AI-governance policy, alongside more technical frameworks like the NIST AI Risk Management Framework and ISO/IEC 42001. Concretely, this typically means: documented accountability for AI systems used in research workflows (from AI-assisted literature review to AI-assisted data analysis); disclosure to affected parties (co-authors, reviewers, study participants, research subjects) when AI materially shaped a research output or decision; and testing/validation proportionate to the AI system’s role, rather than a single blanket sign-off. See also Responsible AI and Trustworthy AI for the related concepts these Principles operationalize, and Generative-AI disclosure statement for the specific mechanism most journals and funders now use to satisfy the transparency principle at the point of publication or proposal submission.
Sources
- OECD — AI principles
- OECD.AI — AI Principles overview
- OECD press release, 8 May 2024 — OECD updates AI Principles
Frequently Asked Questions
Are the OECD AI Principles legally binding?
No. The Principles are set out in the Recommendation of the Council on Artificial Intelligence (OECD/LEGAL/0449), and an OECD Recommendation is not a treaty and carries no legal force of its own. Adherence is a political commitment, with an expectation that adherents will do their best to implement the Principles in national policy and report on that implementation. This is the substantive difference from the EU AI Act, which is binding law with defined penalties.
When were the OECD AI Principles adopted, and when were they updated?
They were adopted by the OECD Council meeting at Ministerial level on 22 May 2019, making them the first intergovernmental standard on artificial intelligence, and were subsequently reflected in the G20 AI Principles adopted later that year. The Council, again at Ministerial level, revised them on 3 May 2024.
What actually changed in the 2024 update?
The 2024 revision responded to the arrival of general-purpose and generative AI rather than rewriting the framework. Several principle and recommendation headings were expanded for clarity; the human-rights principle was extended to cover AI-amplified mis- and disinformation while respecting freedom of expression; language on privacy, intellectual property and safety was sharpened for generative-AI-specific risks; and the text on traceability and risk management was elaborated and moved into the Accountability principle. The count of five values-based principles and five recommendations did not change.
How many countries adhere to the OECD AI Principles?
47 countries and jurisdictions adhere as of the 2024 update. That covers all OECD member states and the European Union, plus non-member adherents including Argentina, Brazil, Egypt, Malta, Peru, Romania, Singapore and Ukraine. Adherence is open to non-members, which is part of why the Principles function as a common governance vocabulary well beyond the OECD’s own membership.
What is the difference between the five principles and the five recommendations?
They address different audiences. The five values-based principles are directed at AI actors — the organisations and individuals who design, develop, deploy or operate AI systems — and describe how those systems should behave across their lifecycle. The five recommendations are directed at national governments and describe what states should do to build a trustworthy AI ecosystem: fund R&D, build the enabling ecosystem and governance environment, prepare for labour-market transition, and co-operate internationally. A research institution reads the first set as operational guidance and the second as context for the policy environment it will be regulated in.
Do the Principles say anything specific about AI in research?
Not directly. The Principles are deliberately sector-neutral: there is no research-specific chapter, and no carve-out for academic or non-commercial use. A university deploying an AI system therefore sits within scope as an AI actor in the same way any other deployer does. What this means in practice for research organisations — accountability, disclosure and proportionate validation — is covered in the section above, and the operational detail is usually supplied by more technical frameworks such as the NIST AI Risk Management Framework or ISO/IEC 42001, which are designed to be implemented against principles like these.
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="OECD AI Principles"
vocab-term-identifier="https://casrai.org/dictionary/term/oecd-ai-principles" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/oecd-ai-principles",
"name": "OECD AI Principles",
"identifier": "https://casrai.org/dictionary/term/oecd-ai-principles",
"description": "The OECD AI Principles are the first intergovernmental standard on artificial intelligence, adopted by OECD member and partner governments in May 2019 and updated in May 2024 to address general-purpose and generative AI. They consist of five values-based principles for trustworthy AI (inclusive growth/sustainable development/well-being; human rights, democratic values and fairness including privacy; transparency and explainability; robustness, security and safety; and accountability) plus five recommendations for policymakers (invest in AI research and development; foster an inclusive AI-enabling ecosystem; shape an enabling, interoperable governance and policy environment; build human capacity and prepare for labour-market transition; and pursue international co-operation for trustworthy AI). As of the 2024 update, 47 countries and jurisdictions adhere to the Principles, including all OECD members, the European Union, and several non-member states. A research institution or research-performing organization can treat the Principles as an operational baseline for AI governance when: (1) it can point to a documented AI risk-management or oversight process addressing all five values-based principles (not just one, e.g. privacy) as applied to a specific AI use case in the research lifecycle; (2) that process is proportionate to the AI system's stage and context of use rather than a single one-time sign-off; and (3) it is paired with actual transparency to affected parties (participants, authors, reviewers) about where and how AI was used. Simply having an internal 'AI policy' document does not, on its own, satisfy the Principles — the OECD frames them as principles for actors across the AI system lifecycle, not a checklist to file away.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/compliance-regulatory#set",
"url": "https://casrai.org/dictionary/term/oecd-ai-principles",
"sameAs": [],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-07-23T04:35:28",
"dateModified": "2026-09-04T07:13:58",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
