Two-year trial: Banning AI in school makes students worse, not better

A law professor has for two years followed three groups of students: one without AI access, one with free and unguided AI access, and one with structured AI training.

The result repeated itself both years: the group without AI did the worst.

What's interesting is that the researcher went into the study expecting the opposite, namely that free AI use would harm the students most. That's not how it turned out.

For you as a parent or teacher, this means: a total AI ban isn't automatically the responsible solution. In practice, it can leave students worse off than giving them proper, structured training in using the tools.

For companies, the point is the same in miniature: banning AI among employees doesn't solve the problem of incorrect use; it just postpones it.

You don't need to change anything today. But if your school, your team or your company has an "AI is forbidden" principle, this is a concrete pointer for starting the conversation: is the ban a protection, or is it just an excuse for not investing in proper training?