A new AI tool lands in your inbox most weeks now, usually with a pitch and almost never with a manual. The question underneath each one is the same: Does this belong in my lesson, and if it does, how? In June 2026, the federal government released Canada’s national AI strategy, AI for All. The strategy promises free AI literacy training and more than doubles the number of K to 12 teachers it aims to reach. It describes literacy as being able to recognize AI, judge what it produces, and decide where it belongs. That last part is a teaching issue before it is a policy one. Deciding where AI belongs is daily classroom work, and it needs a structure.
BEARA provides a structure. Through years of classroom-based work, Johanathan developed BEARA with Mohamed Kharbach, naming it for five steps: Build, Evaluate, Apply, Reflect, Adapt. BEARA is a spiral of five interconnected steps. Each time teachers revisit the steps, they do so with new evidence, deeper understanding, and changing classroom conditions.
Other AI frameworks, including UNESCO’s AI Competency Framework for Teachers and similar guides from groups like Digital Promise, map what capable teachers should know about AI. These maps are useful; however, BEARA is different. It is a route that begins from the moment the tool is open and the lesson is about to start.
The Five Steps
1. Build comes first, because you can’t judge a tool you don’t understand. It means learning enough about how AI works and where it fails to talk about it honestly in your subject area. The question to hold is plain: What do my learners actually need to understand here?
2. Evaluate begins before the tool reaches students and continues during its use. You examine the tool, check what it produces, and watch how it affects learning, participation, privacy, and fairness.
3. Apply is design, not novelty. You build the tool into the lesson for a clear instructional purpose and ask how its use serves the goal.
4. Reflect looks past whether the activity ran smoothly to what happened in student thinking, and whether that was the change you wanted.
5. Adapt continues the spiral. You revise the task or its AI rule based on what you observed, then carry that learning into the next lesson.
Before using AI in a lesson, the routine can be reduced to five questions:
What do students need to understand?
What risks does this tool create?
What learning purpose does it serve?
What changed in student thinking?
What should I alter next time?
What It Looks Like in a Classroom
Picture a teacher who wants an AI tool to help give students feedback on writing drafts before a stack of thirty papers lands on their desk. Build means learning enough about the tool to tell students, honestly, that it predicts likely wording and doesn’t understand their writing. Evaluate means a quick test: feed it one draft with strong ideas and rough sentences, and another with smooth sentences and a weak argument. When the tool praises the polished, weaker piece, you’ve found where it stops being useful. You let it comment on structure and clarity, and you keep the feedback on reasoning for yourself. Apply turns the tool into an exercise: students mark which comments they agree with and which they reject, providing a reason for each, before revising. Reflect can surface something uncomfortable. The drafts improve, then start to sound alike, drifting toward whatever the tool rewards, while work on unfamiliar tasks doesn’t move. Adapt responds: students rebuild their own reasoning before reading the AI comments, and a weekly task with no AI checks whether their thinking carries over.
We find this loop guards against two common failures. Teachers who skip Build go hunting for the right tool with no way to judge any of them. Teachers who treat AI as a one-time checklist tend to drop a tool after one bad lesson, or keep it without question after one good one. Equity belongs in the loop too. When part of a lesson runs through an AI system, a student who can’t access the tool at home or whose language it serves poorly is shut out of how the lesson works. The promise of AI for all only holds in your room if you design for that gap, and that’s what Adapt is for.
There’s a student side to this as well. Teachers can also adapt the five steps to help students think about their own AI use: what the tool is, whether it fits the task, how to use it for a reason, what it did to their thinking, and what to change next time. A Grade 6 student who can say why she turned down an AI suggestion is doing the work the strategy has in mind when it asks Canadians to judge what these systems produce. Handing students that language is part of what makes the practice stick; a rule about when they may log in does less on its own.
You don’t need to teach students how a model is built. What it takes is being willing to make a call about a tool, and to make a better one the next time the lesson runs. If Canada’s AI strategy is going to reach classrooms meaningfully, teachers need routines like this. Small enough to use on Monday and strong enough to guide professional judgement over time. Teachers, schools, or districts interested in trying BEARA are invited to contact the authors for teacher-facing framework materials and to share feedback from practice. BEARA names the steps so the decisions are deliberate. That, more than any single tool, is what AI literacy looks like in practice.
ABOUT THE AUTHORS
Johanathan Woodworth
Johanathan Woodworth is an Assistant Professor in the Faculty of Education at Mount Saint Vincent University in Halifax and a co-creator of the two openly licensed frameworks that the article explains: BEARA, with Med Kharbach, and VAAI, with Med Kharbach and Christine Doe.
Andrea Fraser
Andrea Fraser is an Assistant Professor in the Faculty of Education at Mount Saint Vincent University and spent 18 years in K to 12 education as a teacher, literacy consultant, and literacy coordinator.
This article is featured in Canadian Teacher Magazine’s Fall 2026 issue.
