Responsible Use

The Principle

These tools draft. You decide. Everything on this page follows from that.

UDL Architect, Wonder Web, CrossLink, and REAL Connections generate instructional options for a professional to evaluate. None of them is built to reach a student directly. None of them knows your room, your class, or the child you are actually planning for. What arrives on your screen is a draft written by something that has never met your students, and the distance between that draft and a lesson is your expertise.

Where AI Must Not Decide

Some decisions belong to a human being who knows the child and can be held accountable for the judgment. These tools are not built for them, and should not be used for them.

This is not a limitation of these particular tools. It is a line worth holding with any AI system, including the ones your district may adopt with more confidence than the evidence supports.

Before It Reaches a Student

Every output is a draft until you have read it as a professional. Practically, that means checking four things, because these are the four the model gets wrong:

• Accuracy. Content can be confidently wrong. Verify anything factual you did not already know.
• Standards fit. An output may name your standard and still miss what it actually requires. You know the difference; the model is guessing.
• Developmental fit. Grade band is an instruction to the model, not a guarantee. Read it against the children in front of you.
• Local relevance and representation. The model does not know your community, your context, or who is in your room. Examples, names, and contexts that read as neutral to a general-purpose model may not land that way in your class.

If you would not have handed it to a student had a colleague drafted it, do not hand it over because software drafted it.

Student Information

Do not enter a student's name, work, record, or any identifying detail into these tools. Do not enter it into any general AI tool.

The tools never ask for it. They ask about a lesson, an objective, a standard, or a question. Grade band and subject are context, not records.

If Students See the Output

These tools are for planning, not for student use. But a lesson designed with them ends up in front of students, and that raises questions worth answering honestly.

Tell them

When a resource was designed with AI, say so. Students are forming their own relationship with these tools and they are watching how adults handle them. A teacher who quietly uses AI while telling students not to has taught the more memorable lesson.

Let them argue with it

If AI output appears in front of students, the most valuable thing they can do with it is interrogate it. What did it assume? What did it leave out? What could it not have known? A model's output makes a good object of study because it is confident, fluent, and sometimes wrong.

Supervise it

Any student interaction with these tools or their outputs belongs under the direct supervision of a teacher. Nothing here is designed for a child to use alone.

Attribution

If you share or publish a resource that came substantially from one of these tools, a line is enough:

Designed with the Connected Classroom Intelligence Suite.

This is a courtesy rather than a legal demand, and it exists because colleagues deserve to know what a resource came from, especially when they are deciding how much to trust it.

Report What Goes Wrong

If a tool produces something biased, inaccurate, developmentally inappropriate, or simply strange, tell me: ethics@connectedclassroom.org. Which tool, and what you entered if you are comfortable sharing it.

I read every report myself and reports change the instructions behind the tools. You can also ask for a human review of any output you think is wrong, which is me, reading it. I teach full time, so I will not promise a response window I cannot keep.

Grounding

These commitments are not improvised. They align with the ISTE Standards for Educators, particularly Designer and Citizen, and with UNESCO's principles on human agency in AI. The developmental reasoning behind them is set out at Protecting the Developing Mind, with the underlying research at Research Foundations.