New York and Norway ban AI for the youngest children: sensible caution, but the debate can't stop at prohibition

By WTOP News · October 28, 2026.

The New York City school system has banned students from kindergarten through eighth grade (roughly ages 5 to 13) from using artificial intelligence in the classroom. Norway had already banned its use until after seventh grade. That is according to WTOP, which cites Amy Tondreau, an associate professor of literacy at the University of Maryland, Baltimore County (UMBC), who specializes in early childhood literacy.

According to Tondreau, young children are still developing their cognitive pathways, and those skills are built through repetition and persistence. If AI spares them part of that effort, she argues, they lose the chance to develop them. She says the negative impact on them is disproportionate and different from that on adults. She also acknowledges that the long-term effect of tools evolving so quickly is unknown, and calls for "looking at the long term" before introducing them.

It is worth being clear about what the piece does and does not contain. The text refers to "a growing body of research," but does not cite any specific study: what it offers is an expert's assessment and two policy decisions. These are useful data points, but not quantitative proof of harm.

**Our take.** Tondreau's thesis is intuitive and consistent with what we know about learning: productive struggle, making mistakes and trying again, is part of the mechanism by which children learn to read, write and reason. A tool that hands over the answer before the child looks for it can skip exactly that process. Broadly, it is the same debate we already saw with calculators and search engines, only with a far more capable tool that produces the entire piece of work.

That is why caution with the youngest children seems reasonable to us, especially when the expert herself acknowledges the lack of long-term evidence. The burden of proving that something is safe for child development falls on whoever introduces it, not the other way around.

Even so, the ban has its limits. A ban is easy to announce and hard to enforce: children already encounter AI outside the classroom, at home and on their phones. A school ban without AI literacy leaves students without judgment just when they go out into the world. And there is a risk of inequality: better-resourced families will teach their children to use it well on their own, and others will not.

The position we advocate distinguishes by age and by function. In elementary school, prioritize basic skills without shortcuts. Later on, introduce AI in a guided way, as a tutor that asks questions rather than a machine that answers them, with assessment of outcomes. That is where the long-term potential lies: personalized teaching, accessible to everyone and not just those who can afford a private tutor. But that horizon first requires minds capable of judging what the machine says.

The short term will be confusing, with disparate policies across cities and countries. What matters is that these policies are reviewed with data as rigorous studies arrive, and that today's caution does not harden into dogma.