Not All AI Creates Value for Teachers. Some of It Just Creates Output.

I've been doing some learning lately, working through an online AI for business course, and two terms have been sitting with me. Value creation and value destruction.

Value creation is straightforward. A tool, decision, or product makes someone genuinely better off. It solves a real problem, reduces meaningful friction, or enables something that wasn't possible before.

Value destruction is sneakier. It's when something creates the appearance of help while quietly making things worse, adding complexity, eroding skill development, or solving a problem nobody actually had.

What struck me is how much AI currently being used in education fits the second category, not because the technology fails, but because working and fit-for-purpose are not the same thing.

We know teachers are using AI for lesson planning.

I recently asked a faculty member which AI tools pre-service teachers are using during practicum placements. The answer was ChatGPT, Copilot, and similar general-purpose models. These tools are powerful but they weren't built specifically for lesson planning. They weren't designed with pedagogical frameworks in mind. And they don't centre teacher learning and development in the way tools purpose-built for teacher education can.

When a pre-service teacher uses AI to generate a lesson plan they didn't have to think through, we haven't saved them time. We've taken away a learning opportunity. The product looks polished. The pedagogical reasoning never happened.

That's value destruction with excellent user interface.

Value creation in teacher education looks different. It looks like AI that prompts reflection, surfaces considerations a novice teacher might not yet know to ask and builds the kind of pedagogical thinking that transfers across contexts. AI needs to be structured as a scaffold for thinking, not a replacement for it. The difference between a tool that makes the cognitive work of planning visible and repeatable, versus one that simply does the work for you. (See the comments for an example of such a tool).

So, here's what I keep coming back to.

When Faculties of education are making decisions about tools to support student learning in resource-constrained environments, what should the criteria be? How do we evaluate whether a tool truly supports teacher learning and development, or just makes things faster?

Canadian universities and Faculties of Education are navigating real budget pressure right now. Decisions about which tools to adopt require careful consideration of cost, sustainability, alignment with program goals, and long-term impact on teacher development. But I'd argue that when budgets are constrained, fit-for-purpose AI becomes more important, not less. A tool that doesn't serve the actual goal of teacher preparation isn't a bargain at any price.

Teacher preparation is about building capacity that lasts, about rehearsing the thinking rather than just producing the output. That's what value creation actually looks like in teacher education, and it's a meaningfully different standard than efficiency or speed. Fit-for-purpose AI supports that goal. Everything else, however polished, is just output.

I'd be interested to hear from colleagues in teacher education. What are you prioritizing when it comes to AI tools for lesson planning? How are you thinking about fit-for-purpose in a budget-conscious context?