The First Evidence About AI and Learning

Several new studies this week looked beyond adoption and asked what AI changes inside a classroom. The results were promising in some settings and concerning in others, which is exactly why schools need better evidence.

What Happened This Week

SchoolAI earned an ESSA Tier 3 certification after a two-year study showed 28% improvement in critical thinking.¹²

Researchers at NC State and Carnegie Mellon analyzed 1.4 million interactions in AI-powered classrooms and discovered a sticky pattern.³⁴

A new EdWeek survey reveals most young students cannot tell AI-generated content from human work.⁵⁶

Education Week just published a study on what AI does to teachers, not students.⁷⁸

What These Stories Have in Common

AI can improve critical thinking, help teachers notice students who need support, and reduce administrative burden. It can also make weak understanding harder to detect. These outcomes depend on how the tool is designed and what the teacher can see.

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 begin with small pilots that define the learning outcome in advance. They should collect evidence about student reasoning and teacher practice, then expand only when the evidence shows that the tool is improving both.

The right question is not whether AI works in education. It is which designs work, for whom, and under what conditions.