On Wednesday, New York City banned the use of artificial intelligence in its public elementary and middle schools, which count 600,000 children as students. But millions of American children this fall will open school-issued laptops with an artificial intelligence assistant already installed, ready to draft essays, solve math homework problems and polish slide presentations. In most places, no parent was consulted. No classroom teacher asked for it. The decision was made in Silicon Valley — by people who will never meet these kids or anyone else in the communities raising them.
Technology is transforming our nation’s classrooms at an unprecedented pace. A study published by the Rand Corporation last year found that more than half of middle and high schoolers now use AI for school — an increase of 15 percentage points in just the past two years. Yet less than half of school principals say their school or district has any policy on the use of AI at all.
A separate analysis of AI use across 1,300 school districts revealed that roughly 1 in 5 student interactions with these tools involved cheating, self-harm, bullying or other troubling behavior.
Families are not happy. On behalf of my organization, the Century Foundation, Morning Consult conducted a survey of 2,000 voters. The poll asked about six developments in classroom technology, ranging from students using AI tools to complete assignments without learning, to private companies collecting and profiting from students’ data. In all six cases, at least 80% of Americans said they are concerned.
Growing evidence backs the public’s instincts. Researchers at the Massachusetts Institute of Technology found students who rely heavily on large language models like ChatGPT accumulate “cognitive debt,” outsourcing thinking before they have developed the underlying capacity for it. A separate analysis of AI use across 1,300 school districts revealed that roughly 1 in 5 student interactions with these tools involved cheating, self-harm, bullying or other troubling behavior.
Big Tech talks publicly about AI as an opportunity to enhance learning and support educators. But behind closed doors, many executives view schools as a trove of personal data just waiting to be mined. This year, one of the country’s largest ed-tech companies paid $17 million to settle claims that its college planning software — which many students were required to use — secretly tracked their data, including private messages to school counselors, and shared it with third-party advertising firms.
In Washington, President Donald Trump has dismissed these risks and made AI adoption a centerpiece of his education agenda. He signed an executive order directing the Education Department to prioritize AI funding in its grant programs; the White House gathered more than 60 tech companies to pledge AI tools for classrooms (in addition to welcoming teacher-robots into the East Wing); and the administration shuttered the Office of Educational Technology, the only federal office whose job was helping schools use AI tools responsibly.
It is still true that children learn best from humans — not because we’re nostalgic for some pre-AI era, but because decades of learning science says so.
Congress has done little better. In June, the House of Representatives passed a children’s online safety bill only after gutting the the Senate-passed version’s “duty of care” provision, which would require platforms to be designed safely by default so their engagement-maximizing features don’t foreseeably harm kids. Meanwhile, neither major political party has offered a comprehensive national plan for governing AI in K-12 education.
While too many lawmakers race to embrace AI in schools, families are increasingly asking a more basic question: Shouldn’t we figure out the rules before we hand our kids over to unregulated technology?
The Century Foundation’s polling shows a resounding yes. Almost 4 in 5 Americans want clear government standards governing AI in schools. Nearly half (49%) go further and say AI in classrooms should be kept to a minimum. Remarkably, Republicans are more likely to support both new government rules and keeping AI use to a minimum (at 54%, compared with 47% of Democrats). Less than a quarter of voters (23%) favor the current hands-off approach.
None of this means schools should reject new technologies. For example, the American Federation of Teachers recently built a national academy — alongside some of the very companies making these tools — to train teachers to use AI wisely, because educators should master this technology rather than be mastered by it. We are not against innovation. We are against running an uncontrolled experiment on children without parental consent, independent evidence or meaningful guardrails.
There are several areas of commonsense consensus where lawmakers could start. The most obvious: prohibit vendors from selling or sharing students’ personal data, a policy backed by 9 in 10 Americans. Enact a kids’ online safety bill with a “duty of care” provision so the burden of proof sits with billion-dollar companies, not 10-year-olds. Keep screens and AI chatbots out of the earliest grades, where the developmental risks are the greatest, and give educators and parents a genuine seat at the table before districts adopt new technologies. Finally, invest in what families actually want: smaller classes, safe schools and well-paid educators. It is still true that children learn best from humans — not because we’re nostalgic for some pre-AI era, but because decades of learning science says so.
For years, the tech industry’s motto was “move fast and break things.” These companies have now moved into the nation’s classrooms at record speed. The question is no longer whether AI will be part of our children’s education — it is who decides how, on what evidence and in whose interest. Americans have given their answer: Public schools belong to us, not Silicon Valley.
The post The one thing parents agree on about technology in schools: Limit AI. appeared first on MS NOW.
