Education has a familiar problem with new technology: the classroom moves first, and the policy document arrives afterward to describe what already happened. Textbooks, calculators, the internet, learning management systems, each one spread through daily practice years before any ministry, standards body, or UN agency finished deliberating about it. Generative AI is moving at something closer to escape velocity: fast enough that no single institution catches it by matching its speed, only by adding thrust of its own. Digital Learning Week (DLW) 2026 produced a full slate of reports aimed at closing that distance. The most useful thing those documents contain is a clear-eyed measure of how much distance is left, and UNESCO's own Chairs network might be a more reliable instrument for measuring it.
DLW ran in Paris from September 8 to 11, under the theme "Facts | Frictions | Frontiers." More than 25 education ministers signed a Ministerial Statement on sustaining global education as a common good in the age of AI. Alongside it, UNESCO released a discussion paper, "Sustaining education as a common good in the age of AI: the case for deliberative governance," naming six shifts where that public-good character erodes without oversight, and opened six background papers for public comment, on AI sovereignty, age-appropriateness, safety at scale, procurement, cost, and the right to an explanation. The opening keynote, from the Mila Quebec AI Institute's Shingai Manjengwa, was titled "A nudge or a kick: moving toward AI safety and sovereignty in education." Judged by what the week's official outputs actually commit to, they deliver the nudge, a reasonable first stage for something that will take more than one. The policy briefs those six papers are meant to produce won't exist until Digital Learning Week 2027, a full cycle from now. The design assumes there's time to consult, that deliberation still runs faster than adoption, that the field is waiting.
It isn't. Released the same day as the statement, UNESCO's second annual survey of its Chairs network found that 93% of academics believe their students already use generative AI to produce assignments. That's a self-selected, credentialed slice of the field. The direction of travel is still hard to miss. Look at what they want done about it: less than 1% favor a ban, 2% favor no restrictions at all, and nearly everyone else has already landed on the same improvised middle, allowing it for specific assignments, allowing it with disclosure. UNESCO had already published guidance on AI ethics in 2021 and on generative AI in 2023, but neither document told a Chair whether to fail a paper written by ChatGPT. That granular, day-to-day answer converged on its own, well ahead of this week's reports. It isn't new, either: the first edition of the same survey, a year earlier, found two-thirds of institutions already had or were drafting AI guidance, nine in ten Chairs were already using AI themselves, and only 34% reported a positive experience with AI-assisted assessment. Confidence hasn't caught up with practice since, and this year's numbers say it still hasn't.
That convergence isn't as solid as it looks. The Ministerial Statement itself names changes in how learners think as one of the shifts education must reckon with, and the only evidence anyone has for that claim is the Chairs survey: 63% of respondents perceive a decline in students' cognitive capacities. Perceive, not measure. In a week themed "Facts," the number anchoring a ministerial-level claim about the human mind is an unverified belief, with no baseline taken before generative AI existed to compare it against. A consensus this many people reached independently is still worth taking seriously. It has survived two years of real classrooms. That is not the same as being measured.
The reports reach for tools that don't yet move at the same speed as the field. Some are still finding their footing. Some are structurally slower by design. "The Algorithm in the Room," UNESCO's companion volume of think pieces released the same week, closes on the hope that education moves from being subject to these systems to authoring them. The background paper on procurement argues purchasing is an underused lever, and it's correct, but a purchase order is a buyer's veto over someone else's system, still someone else's system. The one titled "Is there a right age for AI in education?" is still open for comment, with feedback due October 15; the Chairs survey suggests classrooms have been answering some version of that question informally for at least two years, just not in a form anyone can cite in a policy brief. Even the paper titled "Safe enough to scale?" treats scale as the destination rather than the real question: whether everything in education needs to grow the same way at all.
Instructure's Learning Strategy team names this same instinct Evolving with AI, one of the eight pillars in this year's Impactful Eight: a capacity institutions keep building as the technology, the classroom, and the research all keep moving, not a problem solved once and closed. The team's seat in UNESCO's Global Education Coalition and its Digital Transformation Collaborative makes it a conduit in both directions: UNESCO's research direction and policy scaffolding move into Instructure's product and practice thinking, and what Instructure sees across real classrooms moves back out through the same channel. UNESCO's Chairs network is doing that work today, one adjusted assignment policy at a time, and Digital Learning Week is how the rest of the field gets to see it clearly. It's a public-private partnership doing the same climb from two vantage points, and each one makes the other easier to see.
About the Author
Principal, Global Learning Ecosystem
Simone Ravaioli is a Recognition Technologist and Credentials Cartographer working across education, technology, and global policy. As Senior Director of Global Learning Innovation at Instructure, he leads strategic initiatives that advance learner-centric recognition systems, credential portability, and data interoperability. He works with global partners to translate complex challenges into practical solutions that connect education and employment through innovation and aligned strategy. He currently serves as Co-chair of the W3C Verifiable Credentials for Education Task Force and Chairperson of the Credential Engine Advisory Group. These roles let him contribute to open standards and infrastructure that make recognition more transparent, portable, and meaningful. He's a strong advocate for framing micro-credentialing through the lens of "Policy as Data." The approach bridges the gap between frameworks and practice, helping learners and workers navigate these systems more easily and equitably.