Should Schools Ban AI in the Classroom?

The nation’s two largest school systems, New York City Public Schools and the Los Angeles Unified School District, recently introduced broad restrictions on student use of artificial intelligence (AI) for the 2026–2027 academic year. New York City’s one-year moratorium bars students through eighth grade from using generative AI tools but permits limited, supervised use in some high school classrooms. Los Angeles temporarily blocked student access to generative AI tools across all grades.

As national debate over AI’s role in society intensifies, other school districts have adopted policies that limit classroom technology. Over the past year, several states have enacted measures to curb students’ cell-phone and social media use.

School districts have adopted these restrictions in response to concerns from parents, teachers, and education professionals about AI’s effects on children. Policymakers also struggle to adapt existing legal frameworks to AI. The technology poses questions involving privacy, data-center construction and its effects on local communities, and the balance between state and federal regulation.

Other districts have taken less restrictive approaches. The Dallas Independent School District promotes responsible AI use by students and provides AI training for educators. The Houston Independent School District, the largest school district in Texas, recently launched an AI-powered learning pilot at two elementary schools.

At the federal level, President Donald Trump and members of Congress have also sought to address AI’s increasing role in education. A 2025 Trump Administration executive order established a White House Task Force on AI Education to promote youth AI literacy. In Congress, a bipartisan group introduced the Recommending Artificial Intelligence Standards in Education Act, which would encourage states to develop standards for AI use in elementary and secondary education. Another bill introduced last year, the Learning Innovation and Family Empowerment with AI Act, seeks to bolster student privacy protections and provide parents with more control over K-12 AI use. Neither bill has become law.

Supporters of limits or bans point to research suggesting that classroom AI may have negative, long-term effects on cognitive development and learning outcomes. Critics contend that excluding AI from classrooms could force students to develop essential AI literacy skills without guidance from educators. Others critics warn that bans could worsen educational inequities. They argue that students with home access will continue to develop AI literacy, but students without home access could fall behind.

In this week’s Saturday Seminar, scholars and policymakers assess the risks and opportunities of classroom AI use and propose ways to maximize its benefits while minimizing its risks.

• In a report published by the Brookings Institution’s Center for Universal Education, Mary Burns and several coauthors assess the role of AI in education. Burns and her coauthors find that adopting classroom technology has not, on its own, improved learning outcomes. In the AI context, the Burns team contends that programs should strengthen rather than replace human knowledge and skills and should protect cognitive and social development. Although the authors conclude that the risks of educational AI outweigh the benefits at this stage, they remain optimistic that AI can deliver meaningful benefits. They recommend establishing comprehensive regulatory frameworks for educational AI and integrating AI literacy into teacher training.
• In an article in Discover Education, Oluwatayo Osodein of Nigeria’s Kogi State University and several coauthors examine whether existing legal and institutional frameworks adequately protect student rights as AI platforms expand throughout the U.S. education system. The Osodein team identifies concerns involving third-party access to student data, commercial exploitation, unclear federal and state responsibilities, and limited transparency in algorithmic decision-making. The authors propose an “Algorithmic Accountability and Data Governance Framework” that would update federal statutes, require independent audits of school algorithms, and give students greater control over their own data. Osodein and his coauthors call for stronger federal standards, reforms to student privacy laws, and meaningful student consent to promote equitable education outcomes.
• In an article in Frontiers in Psychology, Blanka Klimova and Marcel Pikhart, both of Czechia’s University of Hradec Králové, synthesize research on AI’s effects on students’ well-being. They identify potential benefits, including personalized learning and reduced academic stress, along with concerns about loneliness, technostress, privacy, and reduced face-to-face interaction. Klimova and Pikhart warn that excessive reliance on AI may weaken interpersonal skills and emotional intelligence. They recommend a human-centered approach that emphasizes AI literacy, critical evaluation of generated content, privacy safeguards, and further empirical research. The authors emphasize that educational policy should balance technological efficiency with student well-being and meaningful human interaction.
• In a guidance document, the Massachusetts Department of Elementary and Secondary Education explains how local educators should integrate AI in ways that advance the state’s educational mission while also promoting safe, responsible, and equitable technology use. The Department identifies core principles for AI use in Massachusetts schools, including commitment to data privacy and academic integrity. The Department also provides implementation frameworks that districts can use to assess progress toward goals such as student and teacher AI literacy. In one framework, the Department explains how districts can align AI implementation with legal requirements involving privacy, accessibility, data governance, and public accountability.
• In a paper published by the Association for Computing Machinery, Sabrina Hackman of the University of West Florida and several coauthors argue that existing student privacy and online protection laws do not meaningfully regulate AI surveillance in schools. The Hackman team reports that schools increasingly use AI tools to track student behavior and create predictive inferences. The authors contend that the Family Educational Rights and Privacy Act (FERPA) does not cover these predictive outputs, leaving students vulnerable to harm from unregulated surveillance systems. They recommend modernizing FERPA to cover predictive outputs, limiting nonessential use and retention of student data, and creating accountability structures to restrict data vendors’ use of student information.
• In an article in the ECNU Review of Education, Yong Zhao of the University of Kansas argues that concerns about AI in education, although real, obscure a deeper structural problem: “the growing obsolescence of traditional schooling.” The problem, Zhao explains, is not simply that AI enables students to complete tasks once requiring cognitive engagement, but that schools continue to reward automatable work. Zhao posits that the traditional educational model remains organized around “routine production, standardized performance, and narrow proxies for learning.” He recommends reconsidering curricula, pedagogy, learning environments, and assessment to redefine what students should learn and how schools should organize learning in an AI-saturated world.