UNESCO's Work on Education, Learning and Technology
UNESCO treats education as a human right and a public good, and that framing shapes everything it does on technology in the classroom: the goal is human-centred, ethical and equitable use of digital tools, not adoption for its own sake. Much of that work now runs through its digital-education programme, which spans policy guidance, research, and direct partnership with governments and institutions — with particular attention to disadvantaged and low-resource settings, where the gap in access is widest.
The centrepiece of UNESCO's current output is its response to generative AI: a Recommendation on the Ethics of AI, dedicated Guidance on Generative AI in Education and Research, and twin AI competency frameworks for teachers and students that emphasize human agency, critical thinking and ethics over raw technical adoption. That work converges each year at Digital Learning Week, UNESCO's flagship Paris gathering of ministers, researchers and practitioners on digital learning and AI policy.
Digital Learning Week, launched in 2023, has grown into one of the largest recurring events in the field — the 2025 edition drew more than 400 participants on-site under the theme "AI and the Future of Education: Disruptions, Dilemmas and Directions," and produced a 160-page global report of the same name, examining AI's role in teaching, assessment and workforce development alongside its risks. The 2026 edition, running 8–11 September at UNESCO Headquarters in Paris under the theme "Facts. Frictions. Frontiers. Education in the age of AI," is expected to bring together roughly 1,200 in-person participants and 3,000 more online, including around 30 ministers of education.
Underneath the AI-specific work sits UNESCO's longer-running mandate on digital learning policy more broadly: supporting Member States in building national digital-education strategies, maintaining gateways to digital public learning resources, and — through bodies such as the Institute for Information Technologies in Education (IITE) — running research and capacity-building programmes aimed squarely at the countries least likely to benefit automatically from new technology. The throughline across all of it is UNESCO's own stated position: technology should serve the Education 2030 Agenda and its Sustainable Development Goal on inclusive, equitable quality education — not the other way around.
OECD's Work on Education, Learning and Technology
Where UNESCO frames education technology as a rights and equity question, the OECD approaches it as an evidence and measurement question: what actually happens to learning when a given technology enters the classroom, and how do you know. That work sits mainly with the OECD Directorate for Education and Skills and its Centre for Educational Research and Innovation (CERI), and it's built around two long-running instruments — PISA, the international student assessment, and TALIS, its companion survey of teachers.
The current focus is generative AI. The OECD Digital Education Outlook 2026, released in January 2026, pairs classroom research with adoption data from member countries and lands on a single argument: whether GenAI helps students learn depends entirely on pedagogical design, not on whether it's switched on. The report cites research in which students completed tasks 48% more successfully using generative AI, yet performed 17% worse than a control group once the tool was taken away — evidence, the OECD argues, that output and learning are not the same thing.
That caution is echoed by OECD's own teacher data: drawing on TALIS, the Outlook reports that 37% of lower-secondary teachers already used AI in their work in 2024, and 57% say it helps them write or improve lesson plans — but 72% are concerned it lets students pass off AI-generated work as their own. The OECD's reading of this is that adoption is outpacing guidance, and that the design of a tool, and the teaching around it, matters more than the technology itself.
Looking ahead, the OECD is building AI directly into how it measures learning. PISA 2029 will introduce a new Media and Artificial Intelligence Literacy (MAIL) assessment domain, developed jointly with the European Commission, testing whether students can engage with, create with, manage and critically evaluate AI systems rather than simply use them. That sits alongside the OECD's longer-running Future of Education and Skills 2030/2040 project, which tries to define the broader competencies — not just digital ones — that students will need well beyond any single wave of classroom technology.
The World Bank's Work on Education, Learning and Technology
Where UNESCO and OECD approach ed-tech through ethics and evidence, the World Bank approaches it as a financing and implementation problem: how do you get working infrastructure, content and teacher capacity into systems that don't have them, at a scale that holds up. As of 2025, over 94% of the Bank's education projects included an EdTech component, typically funding digital infrastructure in schools, learning-management platforms, and digital skills training for teachers and students.
To make that investment comparable across countries, the Bank maintains a Global EdTech Readiness Index (ETRI), scoring national systems across six pillars: school management, teachers, students, devices, connectivity, and digital resources. Its own position is deliberately unglamorous about technology itself — EdTech is not a silver bullet, and without safeguards, AI risks deepening educational inequality rather than closing it.
On generative AI specifically, the Bank co-convenes an AI in Education Global Community of Interest with EdTech Hub, a UK-based research partner, bringing together policymakers, ministries of education and practitioners from dozens of countries to compare what's actually working. Early sessions have drawn officials from almost two dozen ministries of education at a time, focused on questions like how AI tutors change the role of the teacher and where AI-driven personalization helps versus where it risks widening gaps between well-resourced and under-resourced students.
The Bank is explicit that technology risks compounding a different kind of divide: not just who has access to AI tools, but who also has a teacher available to help make use of them — leaving open the possibility that under-resourced students get machine-based learning alone while better-resourced peers get both the technology and the teacher-led support to use it well.
ITU and UNICEF's Work on Education, Learning and Technology
Neither the ITU nor UNICEF are education bodies first — the ITU is the UN's telecommunications agency, and UNICEF's mandate is children's welfare broadly — but together they run the infrastructure layer that everything above depends on. Their joint Giga initiative works with governments to connect every school to the internet and generate the data needed to guide investment, with a target of universal school connectivity by 2030.
Connectivity alone doesn't produce transformation, though, and the two agencies address that gap jointly with UNESCO through the Gateways initiative, which helps countries build interoperable digital education systems — so that platforms, data and services actually work together, rather than each ministry or donor project standing up its own disconnected tool.
UNICEF's own portfolio extends well beyond connectivity: its Global Learning Innovation Hub runs a wide slate of programmes, including the Learning Passport, a portable digital-learning platform originally built for displaced and refugee children, and the Global AI for Learning Alliance (GAILA), which explores how AI tools can support learning in low-resource settings specifically.
The division of labour that emerges across these three UN-system players is fairly clean in practice: the ITU worries about whether the school has a usable internet connection at all, UNICEF worries about whether the resulting platforms actually serve children's learning and protection needs, and UNESCO — through Gateways — worries about whether the resulting patchwork of tools can function as a coherent system rather than a pile of incompatible pilots.
The European Commission's Work on Education, Learning and Technology
The EU runs a parallel track to OECD and UNESCO's global work, scoped specifically to its member states. The Digital Education Action Plan (2021–2027) sets out a common vision of high-quality, inclusive and accessible digital education across Europe, organized around two priorities: building a high-performing digital-education ecosystem, and strengthening digital skills and competences for citizens and businesses more broadly.
Notably, this isn't fully separate from OECD's work — the AI Literacy Framework for primary and secondary education is a joint initiative of the European Commission and OECD, feeding directly into the new PISA 2029 assessment domain on Media and AI Literacy. In March 2026 the Commission also published four sets of guidelines for teachers, covering the ethical use of AI and data, tackling disinformation and building digital literacy, making informed choices on digital content, and teaching high-quality informatics.
Much of the coordination work happens through the European Digital Education Hub, an online community launched in June 2022 to address fragmentation in digital-education policy, research and implementation across the EU. By January 2025 it had grown to more than 5,500 members — school educators, teacher trainers, policymakers — and produced over 100 publications and 200 activities.
The Action Plan itself is due to feed into a broader 2030 Roadmap on the future of digital education and skills, a key deliverable of the Commission's "Union of Skills" priority, which will sit alongside related strategies on basic skills and STEM education. The throughline is consistent with the rest of the EU's digital agenda: digital education isn't treated as a standalone initiative but as one strand of a wider push toward a "Europe fit for the Digital Age."
The World Economic Forum's Work on Education, Learning and Technology
WEF's role is different in kind from the others on this list: it's a public-private convening body, not a research institute, a UN agency, or a financing arm, so its output looks less like policy guidance and more like a shared framework for institutions and companies to assess themselves against. Its June 2026 report, "Shaping the Future of Learning: Education Readiness for the Age of AI," introduces an AI Readiness Framework spanning governance, infrastructure and pedagogy.
The framework's central argument is that AI readiness is a precondition for benefit, not a byproduct of adoption: without the right foundations in place, AI can weaken the learning, equity and trust that education depends on, rather than strengthen them. It's written to be used by policymakers, educators, institutions, technology providers and communities alike, to judge whether the conditions for responsible adoption actually exist before rolling AI out.
A recurring point in the report is that the binding constraint on AI readiness differs by country, even when the same framework is applied. In the United States, for instance, where AI tools already spread quickly, the report locates the weakness less in access to the technology and more in the support available for putting it to genuine pedagogical use — echoing the OECD's own finding that adoption without design guidance tends to raise output without raising learning.
That framing puts the World Economic Forum in a slightly different lane from the multilateral organizations above: it isn't building assessments like PISA, financing infrastructure like the World Bank, or setting ethics recommendations like UNESCO — it's giving governments, schools and technology providers a shared checklist to judge their own starting position before any of that other work lands on the ground.
Regional Development Banks' Work on Education, Learning and Technology
Alongside the World Bank's global reach, regional development banks run structurally similar but geographically focused programmes. The Asian Development Bank (ADB) supports digital learning across Asia and the Pacific through financing, knowledge and partnerships, and — much like the World Bank's ETRI — has commissioned diagnostic work assessing "EdTech Readiness" in specific countries, including a technical-assistance programme piloting EdTech interventions across Cambodia, Indonesia, the Philippines and Vietnam based on that diagnostic.
The Inter-American Development Bank (IDB) plays a parallel role across Latin America and the Caribbean, and works comparatively as much as it finances directly — studying other regions' digital-transformation programmes for lessons transferable to its own.
A recent IDB blog, for instance, draws three lessons for Latin American and Caribbean countries from South Korea's digital-education transformation: by December 2024 Korea had achieved over 100% distribution of digital devices for target grades, backed by a $70 million investment in high-speed school internet infrastructure, and was aiming to train 300,000 teachers — including 34,000 "Leading Teachers" — by 2026, alongside a $1 billion investment in AI-powered digital textbooks.
ADB's broader digital-technology work follows a similar logic to its education-specific programmes: it pairs connectivity investment — extending internet access to schools and underserved communities — with digital-skills training and policy support, on the view that infrastructure and capability have to be built together rather than infrastructure first and skills later.