I’ve been following AI in education closely since November 2022. I’ve written about it a lot, probably more than most people want to read. So when UNESCO released its 2025 volume AI and the Future of Education: Disruptions, Dilemmas and Directions, I wanted to see where the global conversation has landed. This report confirmed some of what I’ve been arguing for a while now. It also pushed my thinking in a few unexpected directions.
Let me start with what I found most compelling. The report brings together 21 think pieces from scholars across philosophy, AI research, pedagogy, and policy. The editors frame AI in education as a historical rupture, a moment of genuine tension between the speed and scale of commercial AI systems and the slower, relational rhythms that make education meaningful. “These conversations engage with urgent and unresolved questions about the nature of intelligence, the purpose of education and the futures we are co-constructing, consciously or otherwise, through our entanglements with AI” (p. 14).
The Tool Trap, Revisited
One of the arguments I’ve been making for some time is that using AI effectively in teaching requires more than collecting tools. I’ve seen the “tool trap” before play out across social media and professional development sessions. My roundup posts of AI tools consistently get more engagement than anything else I share online. I get it. Tools feel concrete. They give you something to try on Monday morning. But stacking tools without pedagogical grounding leads to what AI workslop, polished-looking but substance-thin outputs that give the impression of innovation without delivering real learning.
UNESCO’s report echoes this concern from a different angle. Several contributors insist that AI integration must begin with pedagogy. Human relationships anchor education. AI can support that relational work, but it can’t replace it. That aligns with the framework I developed with Dr. Jonathan Woodworth, BEARA (Build, Evaluate, Apply, Reflect, Adapt), which positions tools as only one component of a much larger professional process. The “Build” domain in BEARA, where teachers develop foundational AI literacy across knowledge, skills, and attitudes, comes before any tool selection. The UNESCO volume reinforces why that sequence matters. You need to understand what AI does and doesn’t do before you can make informed decisions about how to use it in your classroom.
Efficiency Is Not the Goal
Here’s where the report pushed me to think harder. Several philosophical essays challenge the very idea of treating AI as an efficiency tool. Báyò Akómoláfé, for example, questions the assumption that learning is individual and cognitive. He frames it as relational, ecological, and more-than-human. That’s a very different conversation from the one we usually have in schools. And it forced me to ask myself something uncomfortable: when I encourage teachers to “try AI,” am I sometimes defaulting to efficiency arguments? Save time on lesson planning. Generate rubrics faster. Produce feedback at scale.
I still believe those are legitimate uses. But the UNESCO volume reminded me that efficiency should never be the primary justification for anything we do in education. Education has purposes that go beyond speed: developing judgment, cultivating autonomy, building the capacity to think things through when the answer isn’t obvious. Emily Bender puts it plainly: large language models “do not understand, reason or care; they produce statistically plausible text without intent or meaning” (UNESCO, 2025, p. 15). That doesn’t mean AI has no place in education. It means we need to be clear-eyed about what these tools actually do so we can position them appropriately in our teaching.
Assessment Under Pressure
The section on assessment hit close to home. Generative AI can produce polished assignments. Traditional assessment models are struggling to keep up, and everyone knows it. The volume maps out this tension without settling on a single answer. Some contributors propose graduated frameworks to guide AI use in coursework. Others imagine continuous, formative, human-centered assessment supported by AI feedback systems. The editors warn that “the future of assessment risks becoming another site of exclusion” (p. 16).
I agree with that warning. And it connects to something I’ve observed in my own work with graduate students. When assessment stays fixed on final products, AI becomes a shortcut. But when we design assessments around process, reflection, and iteration, AI becomes a thinking partner. That’s the difference.
Equity Cannot Be an Afterthought
The report makes a strong case that AI in education is fundamentally an equity issue. One-third of humanity remains offline. Access to advanced AI models often requires infrastructure, subscriptions, and linguistic privilege. If AI-powered education becomes the norm, who gets left behind? And who benefits most?
I write about AI tools and strategies every week on my blog, and I’m aware that my audience is largely teachers who already have access to devices, internet, and institutional support. But a huge portion of the global teaching workforce doesn’t share those conditions. UNESCO is right to insist that governance and policy must address these gaps before they harden into permanent structures. AI integration without equity consideration is just another layer of exclusion wrapped in the language of progress.
Pedagogy First, Always
The report’s concluding message resonated with me: “the stories we tell about AI and the futures of education are still being written, and that we have a collective responsibility to shape these stories with care, clarity and courage” (p. 19). I read that as a call to action for every teacher, researcher, and education leader who cares about this work.
My position hasn’t changed. I still encourage teachers to step outside their comfort zone and experiment with AI. I still believe avoidance is a losing strategy. The teachers who engage with this technology now, critically and intentionally, will be better prepared to serve their students in the years ahead. But this UNESCO volume strengthened my conviction that the “how” of AI integration matters enormously. You can use AI in your teaching and still miss the point entirely if you treat it as a plug-in for productivity. The real work is building an AI pedagogy, one grounded in your disciplinary knowledge, attentive to your students’ needs, and clear-eyed about what these tools can and can’t do.
That’s the project. And it’s far from finished.
Reference
UNESCO. (2025). AI and the future of education: Disruptions, dilemmas and directions. UNESCO. https://doi.org/10.54675/KECK1261