Over the past six months, I’ve had conversations with educators across the globe about AI. How they’re using it, how and why they’re not using it, and how they’re thinking about it.
Those conversations have been incredibly important to understanding the perceptions around the value of AI in the education space. But they’ve also been oddly consistent. The questions keep circling the same territory: Which models are safe to use? How is student data protected? What policies do we need? How do we make sure everyone—students, faculty, and staff alike—understands what they’re dealing with?
But there’s a question we’re mostly not asking, and I think it’s the one that matters most: Who is our AI learner right now, and what does good design actually look like for them?
That’s what this series is about. Part one starts with what we already know. Part two will look at what it means to shift from transferring knowledge to building capability. Part three will get into how we design learning experiences that actually prepare students for what’s ahead.
Let’s start with the data, because it’s not subtle.
The numbers aren’t a warning sign; they’re a description of now
Microsoft’s 2025 AI in Education Report found that 93% of US students have used AI for school-related work. That number jumped 26 percentage points in a single year. A separate 2025 survey by Studiosity and YouGov found 82% of US students have turned to AI for assignments or study tasks. The Digital Education Council’s Global AI Student Survey reports that more than half of students use AI weekly, and a quarter use it every day.
That’s an illuminating truth; AI use among students is the baseline now, not the exception.
Meanwhile, the 2026 Lumina Foundation-Gallup State of Higher Education study tells a different story on the institutional side. About half of US college students say their school discourages AI use, and another 11% say it’s outright prohibited. Only about four in 10 say they’re actually encouraged to use it.
That gap is real and it’s growing. Learners have moved on while many institutions haven’t. And while that distance stays open, the chance to teach AI use thoughtfully, based on judgment and genuine skill-building, keeps slipping.
Three shifts worth sitting with
Let’s look at three deeper shifts happening among student AI adoption, and explore what it means to be a learner today.
Behavior. Students aren’t asking permission. They’ve already woven AI into how they read, write, study, and prepare. When we design assignments without acknowledging that, we’re not stopping AI use. We’re just losing visibility into it. The work is still getting done. We’ve just taken ourselves out of the conversation about how.
Expectations. Learners now expect personalization, speed, and relevance as a given. If a textbook explanation doesn’t click, an AI tutor will try five different ways until one does. If a draft needs feedback, they don’t wait for office hours. The bar for instruction has quietly shifted to adaptive, on-demand, and tailored—and anything that feels generic by comparison feels like a step backwards. A degree is no longer the only path learners see as credible. Micro-credentials, bootcamps, and self-directed AI-assisted study are all on the table.
Motivation. This is the deepest one. For a long time, the logic of academic work was clear: do the assignment, earn the grade, get the credential. Effort was evidence of learning. That logic falls apart when AI can complete the assignment in 30 seconds. And students know it. The honest question a lot of them are sitting with is: what’s the point?
The answer can’t be that the assignment matters because it always has. That won’t hold. The answer has to be that the work demonstrates something AI can’t: real capability, sound judgment, the ability to adapt. Those are exactly the things employers say they need most, and they’re what education has to actually deliver.
This is a presence problem, not a technology problem
If 93% of learners use AI and most institutions either ban it or ignore it, what we’re looking at isn’t a technology gap. It’s a presence gap.
We’re still designing learning for a student who doesn’t really exist anymore. Not because learners have changed who they are, but because the context they’re operating in has shifted completely, and our design hasn’t followed.
The work isn’t about banning a tool or bolting one on. It’s about actually seeing the learner in front of us. That takes presence. It takes being willing to sit with what’s true before reaching for a policy or a solution.
In part two, I’ll dig into what that presence makes possible: a genuine rethinking of what education is actually for, when knowledge is everywhere and execution can be automated.
