AIUNESCO's Recommendation on the Ethics of AI: Guiding Principles and Policy Tools
UNESCO got 193 countries to agree on AI ethics principles—no small feat. The 2021 Recommendation covers human rights, peaceful societies, diversity, and environmental protection with ten core principles and eleven policy areas. If you are looking for a global baseline for responsible AI, this is it.
Verified for technical accuracy — Kodi C.
UNESCO got 193 countries to agree on AI ethics principles—no small feat.
Overview
In November 2021 UNESCO’s 193 member states unanimously adopted the Recommendation on the Ethics of Artificial Intelligence, the first global standard-setting instrument on AI ethics. The Recommendation recognizes that AI holds major potential but also threatens human rights, privacy, equality and the environment if left unchecked. UNESCO notes that protecting human rights and dignity is the Recommendation’s cornerstone, and that its foundation rests on principles such as transparency, fairness and the need for human oversight over automated systems. Unlike many aspirational ethics documents, the Recommendation pairs high-level values with detailed policy action areas covering data governance, environment and ecosystems, gender, education and research, health and social well-being and more.
Context and adoption
The Recommendation emerged from UNESCO’s mandate to promote human rights and ethical use of science. After two years of consultations involving governments, academia, industry and civil society, it was adopted at the 41st session of UNESCO’s General Conference on 24 November 2021. The document helps help governments harness AI responsibly and is applicable to all UNESCO member states, offering guidance on balancing innovation with ethical safeguards and protecting fundamental rights in the age of automation. Because it is the first international normative instrument on AI ethics, it fills a regulatory gap and sets up a baseline for national and regional AI policies.
Dynamic definition of AI
UNESCO deliberately interprets AI broadly as any system capable of processing data in ways that resemble intelligent behavior. This dynamic understanding recognizes that technological advances occur rapidly; a narrow definition would quickly become obsolete and make future‑proof policies impossible. By adopting a flexible definition, the Recommendation can accommodate emergent AI techniques such as generative models, neuromorphic computing and edge AI.
Core values
The Recommendation identifies four foundational values that should underpin AI systems. First, AI must respect, protect and promote human rights and human dignity; human beings should remain the ultimate beneficiaries of AI. Second, AI should promote peaceful, just and interconnected societies, ensuring that technologies foster social cohesion. Third, the Recommendation stresses diversity and inclusiveness, requiring that AI systems be designed and deployed to help marginalised communities and avoid reinforcing discrimination. Fourth, AI development must support environmental and ecosystem flourishing, recognizing that sustainability is a precondition for long-term human well-being.
Human rights approach: ten principles
Building on these values, the Recommendation lays out ten core principles for a human-rights-centerd approach to AI. The first principle, proportionality and do no harm, states that AI use must not exceed what needs to achieve a legitimate aim, and risk assessments should prevent undue harms. Second, safety and security requires avoiding unwanted harms and defending against adversarial attacks. Third, the right to privacy and data protection emphasizes strong safeguards across the AI lifecycle. Fourth, multi-stakeholder and adaptive governance calls for inclusive decision making and respect for international law. Fifth, responsibility and accountability mandates auditing, impact assessments and traceability. Sixth, transparency and explainability requires that AI systems be sufficiently open for users to understand decisions. Seventh, human oversight ensures that AI does not displace human responsibility. Eighth, sustainability urges assessment of AI’s environmental impacts. Ninth, awareness and literacy promotes public education and digital skills. Tenth, fairness and non‑discrimination calls for AI actors to advance social justice and avoid perpetuating biases. A UN‑wide set of principles endorsed in September 2022 mirrors UNESCO’s framework, emphasizing do no harm, defined purpose and proportionality, safety and security, fairness and non‑discrimination, sustainability, privacy and data governance, human autonomy and oversight, transparency and explainability, responsibility and accountability, and inclusion and participation.
Policy action areas
To operationalize these principles, the Recommendation identifies eleven policy areas, including data governance, environment and ecosystems, gender, education and research, health and social well‑being, economy and labor, culture, sustainable development, peace and security, multistakeholder participation and international cooperation. For example, governments should develop strong data governance frameworks ensuring quality, security and privacy.
Policies should also foster innovation that benefits the environment and ensures AI supports the United Nations Sustainable Development Goals. Attention to gender equality and diversity aims to avoid worsening existing inequalities, while education policies should promote AI literacy among policymakers, businesses and the public.
Implementation mechanisms
The Recommendation provides concrete tools to translate its principles into practice. Ethical Impact Assessments (EIAs) require teams to examine societal impacts before and during AI deployment. To help countries gauge their readiness, UNESCO developed a Readiness Assessment Methodology (RAM), which assesses a Member State’s preparedness and guides capacity-building support.
Capacity-building initiatives should improve digital literacy and awareness of AI risks and opportunities among all teams. Oversight and monitoring mechanisms, such as independent audit bodies, ensure accountability, transparency and compliance. International cooperation encourages states to share knowledge and good practices to ensure AI benefits are distributed equitably.
Platforms and partnerships
UNESCO has created specialized platforms to advance the Recommendation’s goals. The Women4Ethical AI expert platform brings together 17 leading female experts from academia, civil society, industry and regulatory bodies to share research, promote non-discriminatory algorithms and encourage girls and under-represented groups to participate in AI.
The Business Council for Ethics of AI, co-chaired by Microsoft and Telefónica, fosters collaboration among companies in Latin America to strengthen technical capacities, design Ethical Impact Assessment tools and develop regionally appropriate regulations. These initiatives show how public–private partnerships can amplify ethical guidelines and provide tangible support for practitioners.
Purpose and significance
According to UNESCO, the Recommendation offers guidance to policymakers on balancing innovation with ethical safeguards. It seeks to protect fundamental rights, promote international cooperation, and embed accountability and responsibility into AI systems.
For regions like Northern Ireland, commentators note that adopting the Recommendation helps align local policies with global standards, strengthens human rights protections, provides practical guidance for business and government, and fosters trust and inclusion in AI adoption. The Recommendation also serves as a reference for other international instruments, such as the UN system-wide Principles for the Ethical Use of AI, demonstrating its influence beyond UNESCO.
Wrapping up
UNESCO’s Recommendation on the Ethics of Artificial Intelligence represents a milestone in international AI governance. By linking four core values with ten principles and actionable policy areas, it moves beyond aspirational statements to provide a full framework for trustworthy AI. Its adoption by 193 member states shows a global consensus on the need for human‑centric AI that respects human rights, promotes diversity and sustainability, and ensures accountability and transparency. As governments and teams around the world implement the Recommendation’s mechanisms—ethical impact assessments, data governance, capacity building and independent oversight—the document is likely to shape AI policies and practices for years to come, providing a common language and roadmap for ethical AI development and deployment.
Continue in the AI pillar
Return to the hub for curated research and deep-dive guides.
Latest guides
-
AI Governance Implementation Guide
Operationalise the EU AI Act, ISO/IEC 42001, and U.S. OMB M-24-10 requirements with accountable inventories, controls, and reporting workflows.
-
AI Incident Response and Resilience Guide
Coordinate AI-specific detection, escalation, and regulatory reporting that satisfy EU AI Act serious incident rules, OMB M-24-10 Section 7, and CIRCIA preparation.
-
AI Procurement Governance Guide
Structure AI procurement pipelines with risk-tier screening, contract controls, supplier monitoring, and EU-U.S.-UK compliance evidence.
Coverage intelligence
- Published
- Coverage pillar
- AI
- Source credibility
- 90/100 — high confidence
- Topics
- AI ethics · International policy · AI governance
- Sources cited
- 3 sources (unesdoc.unesco.org, unesco.org)
- Reading time
- 6 min
Comments
Community
We publish only high-quality, respectful contributions. Every submission is reviewed for clarity, sourcing, and safety before it appears here.
No approved comments yet. Add the first perspective.