UNESCO: turning ethics into homework governments actually do
Lesson 3 of 5 in Global Soft Law: OECD, UNESCO, and the G7 Hiroshima Process.
In November 2021, UNESCO’s General Conference adopted the Recommendation on the Ethics of Artificial Intelligence — every member state, by acclamation. That roll call is the headline: it includes China, it included Russia, and it includes most of the Global South, which the OECD’s club of mostly wealthy democracies structurally cannot. The text itself — four core values, ten principles, eleven policy action areas — covers the familiar canon plus ground the OECD treads lightly: environment and ecosystem flourishing, gender equality as a named priority, cultural diversity. (For the principle content, see the ethics module; here we care about what UNESCO built next.)
Because here is UNESCO’s bet, and it is the most interesting one in soft law: a recommendation nobody implements is a press release. So UNESCO shipped implementation tooling with the text — two instruments that turn ethics language into completed questionnaires, scored gaps, and published country reports.
Readiness Assessment Methodology (RAM) — the country-level diagnostic
The RAM is a structured assessment a government runs (with UNESCO support) to answer one question: is this country institutionally ready to govern AI ethically? It probes five dimensions — legal and regulatory, social and cultural, scientific and educational, economic, technical and infrastructural — and produces a gap analysis: no AI strategy? No data-protection authority? No algorithm-literate regulators? It says so, in writing.
Dozens of countries across every region have completed or begun RAMs, and for many in the Global South the RAM report is the first governance artifact of any kind — the document that precedes and shapes the first national AI strategy. That is soft law doing what statutes cannot: building capacity where there is nothing yet to enforce.
Ethical Impact Assessment (EIA) — the project-level tool
The EIA operates one level down: a structured assessment for a specific AI system, aimed at teams procuring or deploying AI — especially in the public sector. It walks through the Recommendation’s values and principles as concrete questions about the system at hand: who is affected, what could go wrong, what mitigations and redress exist.
Think of it as the soft-law ancestor of the impact-assessment obligations now appearing in hard law — the EU AI Act’s fundamental rights impact assessment and the Council of Europe’s HUDERIA methodology are recognizably the same move, made binding.
The Global AI Ethics and Governance Observatory
Launched at the 2024 Global Forum in Slovenia, the Observatory is UNESCO’s answer to OECD.AI: a public platform aggregating RAM findings, country profiles, and ethics resources. Where the OECD observatory maps policies, UNESCO’s maps readiness — complementary lenses, and a mild institutional rivalry that benefits everyone who needs the data.
Global Forums on the Ethics of AI
The Recommendation’s standing conference circuit — recent editions in Kranj, Slovenia (2024) and Bangkok, Thailand (2025). The forums are where RAM results are presented, ministers make commitments, and the implementation machinery gets its periodic political recharge. A recommendation with an annual summit ages very differently from one without.
Key terms: UNESCO Recommendation on the Ethics of AI, readiness assessment methodology, ethical impact assessment
Interactive checkpoint quiz (2 questions) — open this page in a browser to take it.