Schools are beginning to define what teachers should know about artificial intelligence. Competency frameworks commonly include technical understanding, ethics, privacy, critical evaluation, and responsible classroom use.
These foundations matter. But knowing about AI is not the same as knowing how to design learning with it.
The deeper capability teachers need is interaction design: the ability to decide when learners should ask AI for support, compare its responses, challenge its assumptions, revise its output, disclose its role, or continue without it.
AI literacy becomes educational only when it changes the structure of the learning experience.
Competency does not automatically become pedagogy
A teacher may understand that AI can hallucinate, reproduce bias, or weaken privacy. Yet that knowledge does not automatically reveal how to design tomorrow's lesson.
Should students use AI before forming an initial view or afterward? Should they receive one response or compare several? Should the teacher assess the final product, the reasoning process, or the student's ability to critique the machine? When does assistance support learning, and when does it remove the productive difficulty the task was designed to create?
These are interaction questions. They connect tool behavior to learning purpose.
Without this layer, AI competency risks becoming a compliance checklist: teachers can describe the technology and its dangers but still lack a repeatable way to use it pedagogically.
Every AI interaction redistributes cognitive work
Learning tasks contain different forms of cognitive work: recalling, interpreting, generating, comparing, testing, explaining, revising, and judging.
When AI enters the task, it does not simply add assistance. It redistributes that work between the learner and the system.
If AI generates an outline, the learner may gain time for synthesis—or lose practice in structuring ideas. If AI gives immediate feedback, the learner may revise faster—or stop developing error-detection skills. If AI explains a concept, it may improve access—or create an illusion of understanding that has never been tested.
The teacher's role is to decide which cognitive work can be delegated and which work must remain visible, effortful, and owned by the learner.
That is a design responsibility, not a tool preference.
Five modes of educational interaction
Teachers can move beyond generic “AI allowed” or “AI prohibited” rules by designing distinct interaction modes.
1. Question
Students interrogate an AI response for assumptions, missing context, unsupported claims, or ambiguous language.
The learning goal is not receiving information. It is developing disciplined skepticism.
2. Compare
Students compare AI outputs with another model, a source, an expert explanation, peer reasoning, or their own first attempt.
Comparison turns evaluation into an observable practice rather than an abstract warning to “check the answer.”
3. Resist
Students deliberately complete a stage without AI when the struggle itself builds the targeted capability.
Resistance is not anti-technology. It protects the cognitive work that assistance would prematurely remove.
4. Revise
Students improve an AI output using disciplinary standards. They must explain what changed and why.
This mode helps learners distinguish surface fluency from quality.
5. Disengage
Students recognize when AI is inappropriate because of privacy, sensitivity, unreliable context, ethical boundaries, or the need for direct human dialogue.
Knowing when not to use a powerful tool is a mature form of literacy.
Start with the learning function
Before choosing an AI activity, teachers should identify the function of the task.
Is the learner practicing retrieval? Building a mental model? Learning to formulate a problem? Applying a concept in unfamiliar conditions? Developing a voice? Evaluating evidence? Making an ethical judgment?
The same AI behavior can help one function and damage another.
For example, automated explanation may support access when a learner is blocked by language. The same explanation may undermine learning when the objective is to diagnose a difficult text independently.
The design sequence should therefore be:
- Define the capability the learner must develop.
- Identify the cognitive work that produces that capability.
- Decide which work AI may support and which the learner must own.
- Select an interaction mode.
- Make reasoning and revision visible.
- Assess transfer without identical assistance.
This sequence keeps pedagogy in control of technology.
Teacher development needs design rehearsal
Professional development should not stop at tool demonstrations or policy briefings. Teachers need opportunities to rehearse decisions.
They can redesign an existing task under different AI conditions, predict how cognitive work will shift, test the activity with learners, inspect unintended behavior, and revise the structure.
Useful training artifacts include interaction maps, assistance boundaries, sample learner traces, assessment criteria, and post-lesson reflection notes.
Over time, these artifacts become institutional knowledge. Schools learn not only which tools were used, but which interaction designs improved learning under which conditions.
Conclusion
AI literacy frameworks provide an essential foundation. Education must now translate that foundation into the design of real learning experiences.
Teachers do not only need to understand what AI is. They need to decide how an interaction with AI redistributes thought, effort, agency, and responsibility.
The future-ready teacher will not be defined by frequent AI use or confident tool operation. The deeper capability will be designing when learners should question, compare, resist, revise, and disengage.
AI literacy becomes mature when it protects the work through which people learn to think.
Key Takeaways
- AI competency knowledge does not automatically produce effective pedagogy.
- Every AI interaction redistributes cognitive work between learner and system.
- Teachers can design five modes: question, compare, resist, revise, and disengage.
- AI use should begin with the learning function, not the available tool.
- Teacher development needs task redesign and rehearsal, not only demonstrations.
FAQ
What is AI interaction design in education?
It is the deliberate design of when, why, and how learners interact with AI so that assistance supports the intended learning capability.
Why are AI competency checklists insufficient?
They describe important knowledge and principles but may not help teachers decide how AI should change a specific task, sequence, or assessment.
Should students sometimes work without AI?
Yes. When independent struggle is necessary to build the target capability, temporary resistance protects learning rather than rejecting technology.
