In part one of this series, I argued that the learner in front of us has already changed, while most institutions are still designing for someone else. In part two, I argued that the value of education is shifting from delivering knowledge to developing capability. This post is about what that actually looks like in practice.
Three capabilities worth building toward
If capability is the goal, we have to be specific, prioritize, and come to a consensus: which capabilities matter most. Throughout my discussions within the learning community, I’ve found there are three foundational ones that continue to resonate.
Agency. Learners who direct their own learning; they set goals, choose tools, evaluate their own progress, and seek out what they need rather than waiting to be told. These learners build something they'll carry into every role they hold. Agency is a concrete set of habits formed through repeated practice, and these learners yield it.
Judgement. Learners who can interrogate AI output, recognize when an answer is wrong, weigh tradeoffs, and apply their own expertise where AI falls short make better decisions. In a world where getting an answer is easy, knowing whether the answer is right and what to do with it, is the actual skill.
Adaptability. Learners leave prepared for roles defined by continuous reinvention. Evidenced by the World Economic Forum, which calls adaptability the defining workforce trait of the next decade. Every other skill on their list flows downstream from it.
Principle 1: Build agency
In practice, building agency looks like replacing compliance-based assessment with project-based work that learners help define. Give students genuine choice over pathways, tools, and how they demonstrate what they know. It looks like making metacognition visible—requiring reflection on what was learned, not just what was produced.
It also means teaching AI use as a learner skill. When to use it. When not to. How to direct it toward a goal rather than just accepting what it generates. These result in concrete choices we build into the experiences we design.
Principle 2: Develop judgment
To develop judgment, make evaluating AI output a learning outcome. Build assignments where learners must critique, correct, or extend AI-generated work. Anchor that evaluation in disciplinary expertise; learners can't assess what they don't understand, which is exactly where field-specific knowledge becomes essential rather than ornamental.
Use case-based learning, ambiguity, and tradeoff decisions as the core unit of practice. Assess the reasoning, which is sometimes stronger than the correct answer. Ask learners to defend the choice, not just deliver the result. The shift is from measuring what someone produced to measuring how they think.
Principle 3: Cultivate adaptability
Adaptability is built through taking something learned in one context and applying it in an unfamiliar one; it’s transfer. Adaptability is built through designing for unlearning, too: helping learners recognize when a prior method has stopped working and giving them practice in letting it go. A crucial trait in ensuring critical thinking is applied in the age of AI.
Adaptability is championed and sustained through feedback loops that don't stop at graduation. Stackable credentials and lifelong learning pathways aren't perks. They're infrastructure for a workforce that will need to keep reinventing itself.
What has to change at the institutional level
These principles change what happens inside and outside the classroom. But they don't happen without institutional will behind them. Institutions need to move:
- From restriction-first AI policies to ones that guide and integrate learning
- From compliance-based assessment to capability-based evidence
- From the degree as an endpoint to the degree as one stage of a lifelong path
- From faculty as content authorities to faculty as judgment coaches
None of these are small changes. Each one touches governance, faculty development, technology investment, and the relationship an institution maintains with learners over time.
But the alternative of continuing to design for the learner who no longer exists, in service of a credential that no longer tells the whole story, isn't really a strategy. It loses the value of the institution and delays the growth of the learner.
AI changes what learners can do. Learning guides who they become.
