Why AI Policies Need Evidence of Learning

New policies and new AI agents arrived in learning platforms this week. Both developments showed how easy it is to manage access to AI without answering the more important question of whether students are still learning.

What Happened This Week

NYC released its long-awaited AI guidelines for public schools.¹²

When an agentic AI tool called Einstein completed entire college courses inside Canvas, the LMS responded by building its own AI agent for teachers.³⁴

What These Stories Have in Common

Rules about privacy, safety, and acceptable use are necessary, but they do not show what happened during an assignment. A school can have a careful policy and still be unable to tell whether a student understood the work.

Koan starts from a simple premise. The final answer is no longer a reliable measure of learning. Schools need to understand how students reached that answer and whether the work made them more capable.

What Schools Can Do

Schools should pair AI rules with a record of the learning process. Teachers need to see drafts, revisions, questions, and changes in reasoning so they can judge the work with context.

A policy controls where AI can appear. Evidence of process helps a teacher understand what AI actually did to learning.