Research says AI in the classroom hurts younger kids as districts pull back

The pullback

WTOP reported on 28 October 2026 that school districts around the US are trying to figure out how to use artificial intelligence in the classroom — while some of the nation's biggest systems are saying they don't want it there at all, at least for younger kids. New York City has banned AI use in classrooms from kindergarten through eighth grade, and Norway prohibited students from using AI until after seventh grade even earlier.WTOP frames the developing research consensus bluntly: a growing amount of research says the cautious decisions are the right ones.

The developmental argument

The core argument concerns when AI helps and when it replaces needed practice. Amy Tondreau, an associate professor of literacy with expertise in early childhood literacy at the University of Maryland Baltimore County, told WTOP that young children are still growing cognitively: they are developing cognitive pathways and the skills they need.

If children are given AI tools to offload work or take cognitive shortcuts, Tondreau argues, they lose the opportunity to develop the skills they will need as adults. Kids develop those skills through repetition and persistence — and with AI in the mix at a young age, critical thinking skills don't grow the same way. Her bottom line: 'There's an outsized negative impact for them that isn't the same when we're talking about adults using these same tools.'

The policy picture

The district-level bans are a response to both incidents and evidence. The New York City system's K-8 prohibition is among the strictest in the country, and Norway's national rule on pre-seventh-grade use goes further at the national level. These coexist with districts that are integrating AI heavily at high school level for writing feedback, tutoring and coding.

Tondreau's warning applies specifically to the youngest students: the brain is still building the circuitry for attention, persistence and critical thinking, and those circuits are built by doing the work, not by getting answers. Adults, with developed skills, can use AI as a productivity tool without the same developmental risk.

What it means

The AI-in-education story is splitting into two very different markets: tools for adults and older students where speed and access dominate, and tools for young children where the research is now actively against broad adoption. For education-technology vendors selling into elementary grades, the regulatory and reputational headwind just got stronger.

For parents and teachers, the operational lesson is a design question: is this tool giving a child a shortcut that replaces practice, or is it creating practice? The distinction, not the tool category, is what the research says matters most.

Caveats and what to watch

The research base is growing but still young, and studies on AI tutoring for older students show benefits that these findings do not erase. The quoted expert's claims are qualitative reasoning about cognitive development, not a controlled classroom trial, and WTOP's report is a single-outlet summary of the landscape.

Watch whether more districts adopt K-8 bans like New York City, whether states create uniform rules like Norway, and whether ed-tech products for younger kids add practice-preserving designs. Watch for controlled studies testing the repetition-and-persistence claim directly.

The research landscape

WTOP's report sits in a wider research picture: multiple studies and school-district evaluations this year found AI tools in classrooms failing to improve student learning outcomes meaningfully, even as adoption spiked. The headline number often quoted is that school districts are struggling to integrate AI into daily lessons and that major systems are questioning whether these deployments work.

The pushback is not anti-technology across the board. The strongest evidence for AI in education is in tutoring systems for older students where the tool provides immediate feedback and spaced practice and where a human teacher remains in the loop. What the research criticizes is the assumption that exposing children to AI as a general writing or answering tool automatically improves learning — the data says it often does not, and for the youngest students it may be actively counterproductive.

The equity angle

Bans and restricted use also raise equity questions: schools that can afford high-quality tutoring and human support may not need AI, while under-resourced districts hungry for low-cost tools may be the ones most tempted to adopt AI shortcuts for young kids — exactly where the cognitive-development concern is highest.

On the other side, some advocates argue that banning AI outright denies students the chance to learn how to use the tools that will be part of their working lives. The tension between protecting cognitive development and preparing students for an AI-shaped world is at the center of the policy fight, and the NYC and Norway decisions show different regulatory instincts for resolving it.

Caveats

WTOP's report is a single news account; the bans cited (NYC K-8, Norway pre-7th grade) are current policy positions and could change, and the classroom research base remains mixed, with some studies showing benefits for specific older-student use cases.

The quotes reflect one expert's professional judgment, not a controlled experiment. District policies vary widely by state and locality, so the national picture is not uniform.