Why I Built the Assignment Design System

Most AI frameworks in education begin at the wrong end of the problem.

They start with the tool. What can ChatGPT do? What can students automate? Which prompts are useful? Which tools are safe? Those are valid questions, but they are not the first question a secondary classroom needs to answer.

The first question is: what kind of thinking are we actually trying to create?

That is the question that led me to build the Assignment Design System.

The problem I kept seeing

In grades 9-12, AI does not create one generic challenge. It creates a different pressure in every subject:

• In English, students can generate fluent prose before they have a position worth expressing.
• In history, they can assemble neat summaries without ever wrestling with interpretation.
• In science, they can produce polished explanations that outrun their own conceptual understanding.
• In math, they can reach answers without building durable reasoning.

The common failure mode is not "students are cheating." The deeper problem is that classrooms can accidentally reward polished output more than thought.

If that is the structure of the assignment, AI simply exposes it faster.

Why secondary school needs its own frame

K-12 guidance often stays too broad to be useful. Secondary schools sit inside a very specific set of pressures:

• disciplinary depth matters
• academic integrity stakes are high
• college readiness is always in the background
• teachers need subject-specific examples, not generic reassurance

A ninth grade humanities teacher, an AP Biology teacher, and a school leader writing AI guidance are all facing the same technological shift, but not the same design problem.

I wanted a frame that could work across those contexts without flattening them into slogans.

The three moves

1. Wonder

Start with a question, tension, or problem that is worth thinking about even if students have access to strong tools.

This is where rigor lives.

A good Wonder stage creates uncertainty. It gives students a reason to investigate, compare, argue, test, or revise. It makes the assignment harder to outsource because the work has somewhere real to go.

2. Tools

Only after the learning goal is clear do we decide what AI should do.

Sometimes AI should help students brainstorm, generate counterarguments, translate, or test an explanation. Sometimes it should be deliberately constrained. Sometimes it should not be used at all.

The point is not to maximize AI usage. The point is to use tools in ways that sharpen thinking instead of replacing it.

3. Evidence

If students are going to use AI, teachers need better evidence of learning.

That means moving beyond final products alone. We need traces of decision-making, revision, judgment, source comparison, oral defense, design rationale, and reflection on what changed.

Evidence is what closes the loop. It tells us whether the student actually did the intellectual work.

What changed once I used it

The framework became useful to me because it changed design conversations quickly.

Instead of debating AI in the abstract, teachers could ask:

• What is the real intellectual lift here?
• Where, exactly, should a tool help?
• What evidence would convince us the student did the thinking?

That shift matters. It moves a school from AI anxiety to instructional clarity.

It also makes policy conversations better. Once teachers and leaders share a language for the three phases — Wonder, Tools, and Evidence — policy can stop sounding like compliance theater and start sounding like classroom design.

Where it came from

The framework did not emerge from a branding exercise. It came out of practical frustration, repeated redesign work, and research into how students are actually using AI.

My Brown work, including student survey data and follow-up interpretation, sharpened the same pattern I was seeing in schools: the biggest risk was not the tool itself. It was assignments that did not require enough visible thinking in the first place.

The Assignment Design System is my attempt to give teachers and school leaders a usable response to that reality.

What I want the framework to do

I do not want schools to become impressed by AI theater.

I want them to become better at protecting the conditions under which students think, explain, revise, and build judgment.

If the framework is useful, it is because it helps people make better decisions about those conditions:

• teachers designing tomorrow's lesson
• department chairs reviewing assessment patterns
• school leaders translating policy into practice

That is the job.

The tools will keep changing. The design logic cannot be that fragile.