There’s something peculiar about the artificial intelligence debate in schools. The majority of it occurs at the abstract level, such as policy conferences where education officials carefully discuss “balancing opportunity with risk,” think pieces about personalized learning, and opinion pieces alerting readers to the epidemic of cheating. In the meantime, decisions are already being made in real classrooms. After playing an AI-themed game created in collaboration with Amazon, a third-grader in Massachusetts receives a Certificate of Completion. Before they have typed a single word on their new Chromebooks, sixth graders at a public middle school discover Gemini pre-installed and nudging them.
The equipment has arrived. The frameworks aren’t.
The real story lies in that gap between the pace of deployment and the pace of governance. In 2024, UNESCO pointed out that AI systems were being implemented in classrooms “in the absence of regulatory frameworks needed to protect learners and teachers.” Two years have passed since then. In general, it still holds true today.
It’s amazing to see how few of the important decisions were made by those who were most impacted. When Gemini was bundled into devices provided by the school, parents were not consulted. The AI literacy curricula that are currently being taught to teachers were not created by them. The terms were undoubtedly not negotiated by the students. The decisions were made through district technology committees, nonprofit collaborations, and procurement channels—areas that hardly ever make the evening news.
This is significant because the decision-making process is non-procedural. It has to do with values. The phrase “technology on our terms” used by UNESCO sounds almost charming until you consider whose terms are actually influencing what kids see on a daily basis. These are not neutral pedagogical choices when a sixth-grader in Boston uses a chatbot licensed from Anthropic to prepare for standardized tests and a second-grade art class uses Adobe Express. They represent business relationships, priorities, and presumptions that should be examined much more closely than they are at the moment.
Additionally, there is a more profound concern regarding the potential effects of these tools on learning in practice rather than just theory. To be honest, no one knows for sure just yet. As of yet, there is no solid proof that generative AI enhances student performance in any quantifiable manner. It’s more obvious what is lost when a student reaches for a chatbot rather than pondering a challenging question long enough to formulate a response. Laziness is not the issue. It’s more structural in nature: comprehension builds upon itself. How much of a problem you can truly solve depends on the prior knowledge you bring to it. AI tools reveal that logic more clearly than before, but they do not alter it.
When applied carefully and purposefully, AI may enable students to accomplish tasks that they would not otherwise be able to. That’s a realistic hope. However, “used carefully and deliberately” calls for things that most schools don’t currently have: qualified instructors, explicit policies, sincere parental involvement, and some accountability for the businesses supplying the technology. As of right now, the used part is far ahead of the careful part.
A definitive ruling on AI is not what American education needs at this time. A genuine discussion about who sits at the table when these decisions are made is necessary. Not on Amazon. Not on Google. Not a vendor’s pitch deck being used by the district’s technology procurement office.
The issue is not whether AI should be used in classrooms. It’s possible that ship has already set sail. The question is whether the public, educators, and parents still have enough influence over how it’s applied to guarantee that the results genuinely benefit kids, not just the businesses that were astute enough to get there first.
