EU Council backs human‑centred AI

School AI policy is starting to grow up. That’s the real story here. Europe just put teachers at the center of its official AI line for education, and a South Carolina district is doing something very similar at the local level. The common thread is simple — schools are getting more serious about using AI, but they’re also getting more serious about saying where it does not get to decide. ### What did the EU actually do? On May 11, the Council of the European Union approved conclusions on “teachers in the era of artificial intelligence.” This is not a new binding law. It’s a political signal from EU governments about how AI should be handled in education. The signal is pretty clear: AI can help, but teachers remain the guides, mentors, and critical decision-makers in the learning process. (consilium.europa.eu) ### Why does “human-centred” matter here? Because the fight is not really about whether AI exists in schools anymore. It already does. The harder question is who stays in charge when AI starts helping with lesson planning, assessment, accessibility tools, and admin work. The EU’s answer is that AI should strengthen teachers’ autonomy and working conditions, not erode them. It also ties AI use to inclusion, fairness, and the well-being of both teachers and learners. (consilium.europa.eu) ### What does the Council want countries to do? Mostly boring-sounding things that matter a lot in practice. National governments are being pushed to improve teachers’ digital and AI skills, support the development of education-specific AI tools, and address unequal access to devices and digital resources. There’s also a strategic angle — the Council text flags Europe’s dependence on AI tools built outside Europe as a problem tied to digital sovereignty. (consilium.europa.eu) ### Is this about banning AI from classrooms? No — basically the opposite. The Council explicitly says AI can improve accessibility for disadvantaged learners, support more individualized teaching, and help schools with administrative tasks. But it treats those benefits as conditional. Teachers need to be able to assess AI outputs critically, explain AI’s limits and biases, and help students think through the social, ethical, and environmental effects of these tools. (data.consilium.europa.eu) ### So why bring up Charleston County? Because the same philosophy is showing up far from Brussels. Charleston County School District’s newly approved guidelines give teachers discretion over generative AI for academic work, but draw a hard line on high-stakes uses. AI cannot be the sole basis for discipline, special-education decisions, placement, or major academic outcomes. That is a very concrete version of “human-centred” policy. (consilium.europa.eu) ### Is the U.S. moving this way too? In pieces, yes. Congress has already held hearings on AI in K-12 education, with a lot of emphasis on teacher training and guardrails. And on May 12, Rep. Randy Fine introduced the K–12 AI Literacy and Readiness Act of 2026, which would give states and districts clearer authority to prepare students and teachers for AI using existing federal resources. Different politics, same direction — schools need rules before AI use gets normalized by default. (europesays.com) ### What’s still missing? Evidence. A 2026 Stanford review says strong causal research on AI in K-12 is still very limited — only 20 papers in its review met that bar. So policymakers are writing rules before there is a deep evidence base on what actually works at scale. That makes teacher judgment even more important, not less. ### Bottom line? (congress.gov) The new consensus is not “put AI in charge of school.” It’s “use AI where it helps, but keep humans accountable when the stakes are real.” Europe just said that out loud at the ministerial level. Charleston County turned it into district policy. More systems are likely to follow. (consilium.europa.eu) (scale.stanford.edu)