Vetting AI tools before they reach the classroom

Welcome to this edTalks coffee. Today we are talking about how schools decide whether an AI tool is safe to put in front of students.

Thank you for having me. In our school we built a vetting process before anyone could use a new tool with a class. The first question is always what happens to student data, where it is stored and who can see it.

So the vetting starts with data rather than with the features?

Exactly. A primary educator on our team wrote a one page checklist. Does the tool need a student login, can we switch off training on our inputs, is there an age rating, and can a teacher see everything a child sees.

What changed once the checklist existed?

Teachers stopped asking for permission one tool at a time. They could check a tool themselves in ten minutes and bring the answers to the digital lead. It made the conversation about evidence, not enthusiasm.

And backward design came into it as well, I think.

Yes. We start from the learning outcome and work backwards. If the outcome does not need the tool, we do not use the tool. Backward design keeps the pedagogy in charge and the technology in service of it.

What would you say to a teacher who is nervous about all of this?

Start small. Pick one lesson, one outcome and one tool that has passed the checklist. Watch what the students do, write down what surprised you, and share it with a colleague the next day.

Thank you. That is a practical place to finish.