For a long time, schools could treat completed work as evidence of student ownership.
Of course, teachers have always known the evidence was imperfect. Every teacher has had the moment where a piece of work arrives a little too polished, a little too fluent, a little too far beyond what the student had shown the day before.
The old question was usually asked with a raised eyebrow and a half-smile:
“Did your mum help you with this?”
It was a familiar classroom moment because the work still felt connected to the learner. Even when there had been help at home, the task usually carried traces of the student’s effort, decisions, understanding and revision. A finished essay, a submitted design, a working prototype, a polished presentation or a completed project gave teachers something visible to assess.
AI changes the relationship between output and ownership. The question is no longer only whether someone helped the student. The deeper question is how clearly the learner remains visible when powerful support is part of the process.
When a student can generate, improve, reorganise or polish the work with AI support, the finished product can no longer carry the whole burden of evidence. Schools need to look more carefully at the learner inside the work: what the student questioned, compared, rejected, revised, justified and finally stood behind.
That is the deeper meaning of Visible Agency.
Learner agency is still the goal. Visible Agency is the evidence. It is the evidence that the learner remained active, thoughtful and responsible when AI was part of the learning process.
This is why the AI conversation in schools needs to become larger than cheating, detection, policy or productivity. Those conversations have their place. Schools need clear expectations, ethical boundaries and shared agreements. But the deeper leadership question is about learning design. Research on generative AI and education increasingly points to the need for frameworks that preserve learner agency, autonomy and meaningful participation as AI becomes more embedded in learning environments (Roe & Perkins, 2024; Viberg et al., 2026).
When AI can generate the output, ownership has to become visible in the decisions that shaped it.
The finished product still matters. Students should still learn to create work that is accurate, thoughtful, coherent, useful and well crafted. Quality remains important. What has changed is the amount of evidence the final product can reasonably carry on its own.
A polished response may show that the student used a useful tool well. It may show that they selected, refined or presented information effectively. It may also hide the most important parts of learning: the uncertainty that was worked through, the judgement that was exercised, the assumptions that were tested, the revisions that were made and the responsibility the student took for the final meaning.
The product remains part of the evidence. The process now needs to carry more of the learning story.
The future of assessment and learning design will require a richer view of student ownership. Teachers will still look at what students produce, but they will also need to see how students arrived there. This aligns with the assessment-for-learning tradition, where evidence of learning is used to guide feedback, next steps and student self-regulation rather than relying only on final performance (Black & Wiliam, 1998; Nicol & Macfarlane-Dick, 2006).
In AI-supported learning, ownership becomes clearer when students can show the thinking that shaped the work. Completion, fluency and polish are useful forms of evidence, but they are incomplete forms of ownership. Ownership becomes stronger when the learner can explain the decisions behind the work.
Many schools value learner agency, and rightly so. Agency is central to meaningful learning. It gives students a stronger role in shaping, directing and taking responsibility for their work.
Agency can become shallow when it is reduced to surface choice. A student chooses a topic. A student chooses a format. A student chooses a tool. A student chooses how to present the work. These choices may support agency, but choice alone does not show whether the learner acted with intention, judgement or responsibility.
Agency becomes stronger when choice is joined to purpose, discernment and ownership. This is consistent with broader understandings of autonomy and self-determination, where meaningful agency involves more than freedom from control; it includes purposeful, self-endorsed action and a sense of competence and responsibility (Ryan & Deci, 2000).
AI makes this more visible because it can support so many parts of the learning process. It can suggest ideas, generate drafts, organise information, provide feedback, create examples, explain concepts, compare arguments and improve presentation quality. That support can be valuable, and it also asks schools to become more precise about where agency lives.
Agency is revealed by whether the student remained meaningfully present in the learning. Did they know what they were trying to understand? Did they ask better questions? Did they evaluate the suggestions they received? Did they recognise what needed to change? Did they reject ideas that did not fit? Did they revise with intention? Did they take responsibility for the final meaning?
These questions move agency from aspiration into evidence. Visible Agency gives schools a way to protect and strengthen the deeper meaning of learner agency in AI-supported learning. It shifts attention from whether AI was present to whether the learner was present.
Visible Agency is a design principle for AI-supported learning. It asks teachers and leaders to design learning so the learner remains visible when AI is part of the process.
The aim is thoughtful visibility: enough evidence to discuss, assess and improve the learning decisions that matter most. Students do not need to produce long process logs or justify every interaction with AI. The task simply needs to make the important learning decisions visible enough to support better learning.
Visible Agency is concerned with the learner’s active role. It looks for evidence of the questions students ask, the choices they make, the judgements they form, the revisions they justify and the ownership they demonstrate. It asks whether students can explain what they did, why they did it, what changed and what they now understand more clearly.
This is why Visible Agency sits at the centre of AI task design. It asks schools to design AI-supported learning with greater clarity about evidence, responsibility and student thinking. The goal is to make the learner visible inside the use of AI.
That protects the dignity of the learning process. It keeps attention on the student rather than the tool. It allows AI to support learning while preserving the human work of questioning, deciding, revising, interpreting and taking responsibility. Recent work on human agency in AI-era education similarly emphasises the need for human oversight, AI-human complementarity, AI competencies and transparency when integrating AI into learning environments (Viberg et al., 2026).
For the broader framework, see Visible Agency: How to Design AI-Supported Learning Without Outsourcing Student Thinking.
Student ownership in AI-supported learning is broader than whether students produced every word, idea, example or structure alone. Ownership means students can stand behind the work.
They can explain the purpose. They can describe the decisions they made. They can show what they accepted, adapted or rejected. They can justify the revisions. They can explain how the work changed and why the final version represents their understanding.
This applies across many kinds of learning. In writing, ownership may be visible in the student’s claim, structure, evidence choices and revision decisions. In design, it may be visible in the user need, trade-off, prototype decision or design constraint. In coding, it may be visible in the logic the student can explain and the fix they can defend. In inquiry, it may be visible in the question the student refined, the source they challenged or the conclusion they adjusted. In a presentation, it may be visible in the message, audience decisions and final explanation.
Ownership is visible when students can explain the thinking that shaped the work.
This is why the existing articles in the Visible Agency series focus on different parts of the same larger argument. Students need to make their thinking visible. They need to protect the learning demand. They need to exercise judgement. They need to take responsibility for their final work. They need to use AI in ways that strengthen metacognition.
Together, they form the practical architecture of Visible Agency.
When AI can do more of the visible work, students need clearer opportunities to show the invisible work of learning.
AI-supported learning asks students to work in a more complex relationship with evidence, support and responsibility.
Judgement becomes essential because students need to evaluate what AI produces. They need to decide whether a response is accurate, useful, appropriate, ethical, relevant and aligned with the learning purpose. They need to compare possibilities rather than simply accept the first fluent answer. AI literacy frameworks commonly emphasise the need to understand, use, evaluate and critically reflect on AI systems and their outputs (Long & Magerko, 2020; Zhang & Magerko, 2025).
Discernment becomes essential because AI can produce responses that sound confident even when they are incomplete, generic or unsuitable. Students need to notice quality. They need to recognise when an answer is plausible but shallow, clear but incomplete, polished but misaligned, useful but requiring verification.
Metacognition becomes essential because students need to understand how their own thinking changed. They need to notice what AI helped them see, what it challenged, what it clarified and what they now understand more deeply. Metacognition and self-regulated learning research emphasises students’ ability to plan, monitor and evaluate their thinking, and to use that awareness to improve learning (Zimmerman, 2002; Education Endowment Foundation, 2025).
These capacities are closely connected. A student who uses AI well does more than produce an improved answer. They evaluate the support, make decisions about its value, revise with purpose and understand how their thinking developed. They are learning how to remain responsible while using a powerful form of support.
This is where student ownership becomes visible. The learner is visible in the question they ask, the suggestion they reject, the comparison they make, the revision they justify and the final position they can defend. This is how ownership becomes visible.
For deeper treatment of these dimensions, see How to Make Student Thinking Visible When AI Is Part of the Process, How to Design AI Tasks That Require Student Judgement, and How to Use AI to Strengthen Metacognition.
Schools understandably begin the AI conversation with questions of use, access and integrity. Did the student use AI? Was it allowed? How much support is acceptable? How do we detect misuse? What should the policy say?
These questions have a place. They help schools create boundaries, clarify expectations and respond to legitimate concerns. But boundaries alone do not design learning. The more important question is what the task requires of the learner.
A task may comply with policy and still leave agency unclear. A student may disclose AI use and still have little to explain about their own thinking. A piece of work may be completed ethically and still provide limited evidence of judgement, metacognition or ownership.
The leadership opportunity is to lift the conversation. Instead of asking only whether AI was used, schools need to ask what learning remained visible. Instead of focusing only on the tool, they need to examine the task. Instead of treating AI use as the centre of the conversation, they need to return to the student.
AI policy matters, but policy cannot carry the whole weight of learning design. Schools also need shared language for the kind of student thinking, judgement and responsibility they expect to see. Visible Agency provides that language.
The stronger questions are design questions.
Where is the learner visible in this task? What decisions must students make? What thinking must students explain? What evidence shows judgement? What evidence shows revision? What evidence shows ownership? What can students defend as their own? Where does AI support the process while preserving the learning demand?
These questions move the conversation from control to design. They help teachers and leaders look beyond whether AI is present and examine whether student agency is visible.
They also create a more productive conversation with students. Instead of simply asking students whether they used AI, teachers can ask them to explain how they used it, what they questioned, what they accepted, what they changed, what they rejected and what they now understand more clearly.
Those questions invite responsibility. They help students see that AI use is an opportunity to practise ownership more deliberately. If students use AI to generate possibilities, they need to judge those possibilities. If they use AI to improve a draft, they need to explain the improvement. If they use AI to test a claim, they need to respond to the challenge. If they use AI to revise, they need to understand the revision.
The stronger question is: Where is the learner visible in the work?
The Visible Agency Design Test gives schools a practical way to examine whether AI-supported tasks still require the learner to be present in the work.
It is designed to help teachers and leaders review tasks before, during or after AI use. The purpose is to sharpen the learning design, not to create another compliance checklist.
The test asks whether a task gives students meaningful opportunities to show:
Each area asks a different version of the same deeper question: can we see the learner inside the work?
Thinking visibility asks whether students have to show the reasoning, questions or decisions behind the product. Judgement requirement asks whether students must evaluate, compare or decide rather than simply accept AI output. Ownership asks whether students can explain and defend the final work. Metacognition asks whether students can describe how their thinking changed. Effort integrity asks whether the task preserves meaningful cognitive effort. Transparency asks whether AI use is clear enough to discuss. Appropriate use asks whether AI is being used in ways that fit the learning purpose.
Together, these areas turn the Visible Agency principle into a practical design conversation. They help schools move from broad concern to shared practice. They give leaders a way to support teams. They give teachers a way to review tasks. They give students a clearer understanding of what responsible AI-supported learning looks like.
The full Design Test should be used as a practical companion to this article. See The Visible Agency Design Test.
Visible Agency has implications at every level of the school.
For leaders, the work is to move the AI conversation beyond compliance alone. Policies and guidelines are necessary, but they need to be supported by a shared understanding of learning design. Leaders can help teams ask better questions about task quality, evidence of thinking, student responsibility and the kinds of learning AI should support.
This is strategic work. It requires schools to build common language around agency, ownership, judgement and visible process. It also requires leaders to protect the deeper purpose of learning in a time when tools can make completion faster and presentation easier.
For teachers, the work is to design tasks where student thinking remains visible. Often, one or two deliberate design moves are enough: asking students to explain a decision, justify a revision, compare alternatives, identify an assumption or defend the final position.
These small moves can change the quality of evidence. They help teachers see how students are thinking, not only what students have produced. They also help students understand that AI-supported learning still asks them to be active, discerning and responsible.
For students, the work is to develop a more mature relationship with support. AI can help them generate ideas, test explanations, receive feedback and improve work. But they remain responsible for the choices they make with that support. They need to learn how to ask better questions, evaluate responses, revise with intention and stand behind the final meaning.
That is a powerful form of agency.
The future of learner agency will be strengthened by designing learning where the learner cannot disappear.
AI invites schools to become more precise about what counts as evidence of learning.
The work students produce still matters. The quality of their explanations, designs, arguments, presentations and solutions still matters. But when AI can help generate the visible product, schools need richer evidence of the learner inside the work.
That evidence appears in the questions students ask, the judgements they make, the revisions they justify, the thinking they explain and the ownership they take for the final meaning.
This is the promise of Visible Agency. It gives schools a way to embrace AI-supported learning while keeping the learner at the centre. It honours the value of the product while strengthening the evidence of process. It helps teachers design learning where support is useful, judgement is necessary and responsibility remains visible.
The future of learner agency will be strengthened by designing learning where the learner cannot disappear.
This article sits within the Visible Agency series.
For the broader framework, read Visible Agency: How to Design AI-Supported Learning Without Outsourcing Student Thinking.
For the visible process side of the work, read How to Make Student Thinking Visible When AI Is Part of the Process.
For the learning demand side of the work, read How to Stop AI From Replacing Student Thinking.
For the judgement side of the work, read How to Design AI Tasks That Require Student Judgement.
For the task ownership side of the work, read How to Design AI-Rich Tasks That Still Require Student Ownership.
For the metacognition side of the work, read How to Use AI to Strengthen Metacognition.
To review AI-supported learning tasks with your team, use The Visible Agency Design Test.