Kelly Clancy said that in October, when her son was in sixth grade at a middle school in Brooklyn, New York, he was given an assignment to create a science experiment and ask Google Gemini, an artificial intelligence chatbot, for feedback.
Mr. Clancy, who has three children in New York City public schools, told teachers that the bot “is just about teaching kids that machines can think.”
Clancy also founded a group called Parents for AI quote in Educational Spaces, which is calling on the city to halt the use of AI in public schools for two years.
The New Yorkers are among a growing number of parent and child development experts across the country who are raising concerns about AI in schools.
In February, more than 1,100 parents in Bend, Oregon, signed a petition asking their local school district to remove generated AI from students’ devices. In April, Fairplay, a national child advocacy group, issued a statement calling for a five-year moratorium on “generative AI products for students” in preschool through 12th grade.
While big tech companies and the Trump administration have encouraged teachers to use AI, arguing that it will help students learn and give them the skills they need to succeed in a technology-enriched world, some parent and child development experts say there is little evidence that AI helps children and may even harm cognitive development.
“There’s an overwhelming sense that edtech companies are deciding what kids learn, and teachers are being put in a technical support role rather than making decisions about what’s best for kids to learn,” academic editor Clancy says.
In March, First Lady Melania Trump convened a White House Summit on Educational Technology. She entered a room with a robot and advocated a world where children could learn from “a humanoid educator named ‘Plato.'”
Meanwhile, Microsoft, OpenAI, and Anthropic have provided millions of dollars in AI training to the American Federation of Teachers, the nation’s second-largest teachers union, the Associated Press reported.
According to a recent survey by National Public Radio and Ipsos, 40% of K-12 teachers said their students use AI in the classroom at least once a week.
MagicSchool, an AI platform for education, has contracts with school districts across the country, including Atlanta, Denver, New York City, and Seattle, to provide tools such as character chatbots that interact with students and chatbots that provide written feedback.
In New York City, “educators are using MagicSchool tools to enhance engagement, differentiation, and instructional effectiveness while maintaining strong instructional practices and community trust,” the company said.
“One of the interesting things about generative AI systems is that you can say, ‘Take this historical moment and provide a correlation to my life today,'” said Amanda Bickerstaff, CEO of AI for Education, which provides AI literacy training to educators.
However, neuroscience and education experts argue that rather than helping students learn, AI could cause “cognitive offload,” or the use of external aids to avoid mental effort.
A study published in the journal Society in 2025 found that people aged 17 to 25 were “more dependent on AI tools and had lower critical thinking scores than older participants.”
Stanford University reported in March that there is little evidence about how AI will impact K-12 education, and it is unclear whether AI is “helping students complete tasks or supporting sustained learning and skill development.”
AI “was never designed to be a learning tool,” said neuroscientist Jared Cooney Horvath, who linked recent declines in test scores nationwide to increased screen time. “The tools that experts use to make their lives easier are not the tools that beginners can use to learn how to become experts. They won’t learn anything by using the same tools.”
Proponents of AI in education argue that the tool is effective in helping students with learning disabilities.
Bickerstaff said dyslexic students who struggle with reading could use AI to translate text into speech, allowing them to better absorb information before writing a paper.
But Horvath points out that such digital tools have been available for decades and do not require generative AI.
“AI is currently being confused with big technology, forgetting that AI is fundamentally new,” Horvath said. “Everything that has worked in the past worked before AI, so the question now is, has AI made it better? And the answer seems to be a clear no.”
Natalie Houston, a therapist and parent of four in Oregon, agrees that AI isn’t helping students. After discovering that third-graders in the Bend-La Pine School District had begun using MagicSchool’s AI tools, she and more than 1,000 other parents signed a letter asking the district to remove the tools from students’ devices.
After parents expressed concerns about students forming bonds with MagicSchool’s chatbot, Raina, the company announced that it had “removed the persona in favor of a neutral AI learning assistant. This change removes the anthropomorphic element.”
“The concerns voiced by parents in Bend La Pine about how chatbots are normalizing unhealthy relationships during a critical period of brain development are valid,” a Magic School spokesperson told Oregon Public Broadcasting in February.
The school board also adopted a resolution in April stating that the district would “develop standards for the use of educational technology, including a list of educational technology applications approved for use in each grade level.” [level]” and remove “non-evidence-based technologies and applications” from student devices.
Other groups are also calling on schools to limit AI. In May, the American Federation of Teachers recommended removing AI tools for students from elementary schools, despite accepting millions of dollars in funding from big tech companies.
This month, more than half of New York City Council members sent an open letter to Mayor Zoran Mamdani and New York City Public Schools Superintendent Kamal Samuels, calling for a two-year moratorium on the use of AI in schools, except for education on the “risks of technology use.”
The city also withdrew plans to open an AI-focused high school following public backlash.
“The previous administration stepped up the gas pedal on AI without real family involvement,” a New York City public schools spokesperson said in an email to the Guardian. “That is not the approach Chancellor Samuels will take. Earlier this year, New York City Public Schools took the first steps to put in place initial guardrails while working with families and communities to develop policies that protect our students. We look forward to sharing more details soon.”
Bickerstaff doesn’t think elementary school students should use generative AI bots, but he thinks a complete moratorium is a bad idea.
Instead, students need training on how AI tools work and their limitations. Bickerstaff said this allows students to “engage more critically, rather than just relying on the tools and being reliant on them.”
She argues that students need to learn how to use AI because companies expect their employees to be “literate for the AI generation.”
“There’s no way to stop students from using it,” Bickerstaff said.
Houston, a parent from Bend, Oregon, said schools shouldn’t need to teach students how to use AI because it is “designed to be intuitive and easy to learn,” and instead focus on building “fundamental academic skills.”
Clancy, a parent in Brooklyn, said teachers have not used AI in her son’s lessons since she raised concerns about the use of Google Gemini.
She predicts the city will approve the moratorium.
Clancy said she would then “give parents in other cities across the country the tools they need to advocate in their own school districts.”
