The Push to Hit Pause on AI in Schools, Before It’s Too Late

One in five student interactions with AI tools on school devices ends in cheating, self-harm, bullying or other harmful behavior. That’s according to real-time data from Securly, a student safety monitoring platform used by thousands of districts. And yet AI is now embedded in nearly every layer of American education, from the software principals use to manage their schools to the tutoring bots sitting in front of kindergarteners.

A growing coalition of parents, researchers and advocacy organizations is sounding the alarm. Led by children’s tech watchdog Fairplay, more than 260 groups and individuals are calling for a five-year pause on generative AI in pre-K–12 schools — not because they think the future can be stopped, but because they believe children shouldn’t have to absorb the cost of figuring out whether it’s safe.

David Monahan, campaign director at Fairplay, says another 800 individual signatures are on the petition alongside the organizational support. The push aligns with a growing body of skeptical research — including a 2024 Pew survey showing that while 25 percent of teachers believe AI does more harm than good in the classroom, only 6 percent are confident the benefits outweigh the risks.

“We’re being told, ‘of course it’s inevitable, of course AI is here to stay,’” says Monahan. “I say that about AI in my own life, in my communities, in my family. But the potential harm is glaring as it relates to education. … We do have the right to speak up and say ‘Who said this was a good idea?’”

The cell phone playbook

The momentum behind the AI pause campaign didn’t come from nowhere, says Monahan. When parents across the country successfully lobbied for cell phone bans during the school day — now law in 23 states, though not yet in Washington — it proved that organized parent pressure could move school policy. Monahan wants to channel that same energy. “That really inspired people to say we can make a difference, not just gently move the needle.”

Now he wants to apply that same pressure to generative AI, calling for a five-year pause on its integration in all pre-K–12 schools until a specific set of criteria is met.

What would it take to make AI safe for schools?

Fairplay’s position statement spans concerns from data privacy to the risk of students offloading cognitive tasks to a bot. It calls for all generative AI products to meet five requirements before entering a classroom:

1. Improve learning outcomes without causing cognitive offloading or impeding human relationships.
2. Demonstrate absolute safety for students. That means addressing issues like addiction, data and privacy risks, exposure to harmful content, risks to mental health and more.
3. Not be used for cheating, academic dishonesty, plagiarism or other unauthorized purposes.
4. Sufficiently consider and prioritize privacy, civil rights, ethics, justice and climate impacts.
5. Never be used in place of teachers, especially for vulnerable populations like neurodivergent students, at-risk students and students of low socioeconomic status.

If products do not meet these standards, Fairplay maintains they should not be used in pre-K–12 schools. According to their press release, “The call for a five-year pause comes as GenAI products are proliferating in pre-K–12 schools. This is despite the fact that AI technology has already proven deadly for young people, including teens who have died by suicide after AI chatbots encouraged them to take their own lives.”

Experts are split on AI

Experts are divided, and even the cautious optimists come with caveats. Priten Soundar-Shah, educator, philosopher, and CEO of Pedagogy Ventures, has written extensively on the subject in Ethical Ed Tech: How Educators Can Lead on AI & Digital Safety in K-12. He’s not anti-AI, but he’s clear-eyed about the risks. Data privacy is his top concern: Without established federal or state regulations, he says, teachers are currently the only buffer between students and potential misuse, and many are still figuring out the tools themselves.

Bias is another red flag. “Feedback is skewed based on perceptions of the child’s race or gender,” he says. “Especially if it’s being used in high-stakes contexts, we should make sure that parents are being informed of it.”

Emily Cherkin, a Seattle-based author, speaker, and prominent voice on children and technology, shares that wariness. “Children’s data is incredibly valuable, especially in the AI era,” Cherkin says. “Schools provide a captive population. This is not just about privacy. It’s about consent and power.”

The parents pushing back

Parents around the country have begun speaking up against integrating AI in schools. “AI integration in schools is an intentional decision that is up to communities. It is absolutely NOT inevitable,” says Shaleen Title, parent and member of Reconnect Malden, a community alliance looking to reduce tech use in Malden, Massachusetts. “Introducing student-facing AI in the classroom requires a significant and proactive commitment to abandon the values that have always been upheld in public education — that the best way to learn is with your peers and well-trained teachers — and replace them with experimental technology.

“Chatbots are clearly risky for children, but Google enables them, and we feel powerless to turn them off. In my district, Gemini is disabled, but parents definitely need to ask about it and make clear that we are not okay with our children interacting with chatbots as part of the curriculum. Jump out of the boiling water.”

Emily Hubbard has seen AI’s classroom creep from multiple angles — as a public school parent, school board member and PhD student in education. The St. Louis mother of four teenagers isn’t impressed. “I don’t love it. I’m against it," she says. Her objections extend beyond the classroom: the environmental toll of data centers, AI’s well-documented hallucinations, and its role in encouraging harmful behavior in young people.

She’s felt the effects personally. When a graduate school peer submitted feedback that was clearly AI-generated, Hubbard says it stung. “It made me feel like I wasn’t worth personal attention. I would have preferred nothing.” In a charter school she observed for her program, she heard a teacher tell students, “You can use Chat for prompts” — an offhand reference that stuck with her. “I was pretty concerned about the informality of that.”

When she raised her concerns with her superintendent and fellow board members — telling them she wanted to ban AI on school devices — the response was a familiar one: Let’s teach them to use it responsibly. She wasn’t satisfied with that answer.

Victoria Livingstone, former professor and mom to a first grader from Summit, New Jersey, stopped teaching “partly because my university students were relying on LLMs so heavily,” an experience she wrote about in TIME. After publication, “I got hundreds of messages from all over the world. It gave me a really good sense of how frustrated teachers are.” Though she thinks AI literacy is an important skill to teach, she believes it shouldn’t replace critical thinking necessary to assess the LLM’s outputs.

How to get involved in your own community

Start by asking questions. Find out how your child’s school district is currently using AI. Talk to teachers, principals, superintendents and board members. You may be told it’s already “baked in” to tools the district has purchased. Push anyway.

Monahan encourages parents to put AI on the agenda at school board meetings, bring other parents along and keep asking until decision-makers engage. For those ready to go further, Fairplay’s action page has the position statement, a petition and a one-pager to share with your community.

Soundar-Shah points out that parents can also demand accountability on data privacy specifically, by asking schools to consult third-party consortiums that set AI standards — independent resources that exist precisely to help districts navigate this.

The window to shape what AI looks like in schools is open right now. “We don’t want kids to be guinea pigs,” says Monahan. The question is whether enough people will speak up before the experiment is already over.