UNESCO's Ethics of AI Recommendation in practice

Three ways to see it

1. 1

   The UNESCO Recommendation is built on eleven principles. Proportionality and do-no-harm. Safety and security. Fairness and non-discrimination. Sustainability. Right to privacy and data protection. Human oversight and determination. Transparency and explainability. Responsibility and accountability. Awareness and literacy. Multi-stakeholder and adaptive governance. Above all of these sit four values: respect for human rights, environmental flourishing, diversity and inclusion, and living in peaceful societies. Read against the Pakistani context the language is striking; the document anticipates a country that is signing it without yet living it.

2. 2

   Way one to use the Recommendation: as a vocabulary. Pakistani bureaucracy speaks in a vocabulary inherited from the 1973 Constitution, PECA 2016, and assorted SROs. UNESCO gives you a shared vocabulary that lines up with the EU AI Act, NIST, OECD, and most international donors. When you write a proposal to the World Bank or AFD, using the UNESCO vocabulary signals competence and unlocks budgets that would not be unlocked by purely local language.

3. 3

   Way two: as the Readiness Assessment Methodology (RAM). UNESCO has published a structured tool to assess a country's institutional, legal, social, scientific, and cultural readiness for AI. Around 50 countries have run it. Pakistan has not. A RAM run on Pakistan would produce a list of structural gaps, but it would also produce, on every page, the leverage points that the next administration could move. Treat RAM less as a report card and more as a road map.

The why-tree

Try this with Claude

AI-edge prompt to try: 'You are a UNESCO RAM facilitator. Ask me 15 diagnostic questions about my organisation across institutional, legal, social, scientific, and cultural readiness for AI. After my answers, score each cluster 0 to 5 and tell me my three weakest cells.' Compare what the model surfaces to what you already suspected.