New York City has joined governments and school systems restricting generative AI use by students, following tighter rules introduced in Europe. Norway and Italy have taken different approaches, while the European Union’s AI Act leaves gaps around student-facing AI tools.

Norway Moved First

Norway banned generative AI in primary schools in June, months before New York City announced its restrictions. The ban took effect at the end of August, covers grades one through seven and allows supervised use up to age 16.

Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram

Prime Minister Jonas Gahr Støre has argued that unrestricted AI could allow children to bypass essential early stages of learning to read, write and perform basic arithmetic.

Italy Chooses Controlled Access

Italy has taken a narrower approach. Under Law 132/2025, children under 14 have needed parental consent to use AI since October 2025. The permission-based model keeps AI available while placing responsibility on parents, contrasting with Norway’s broader restriction.

The EU’s AI Act Leaves A Gap

The EU AI Act classifies certain education uses as high-risk under Annex III, including AI used for admissions, exam monitoring and learning assessment.

However, the designation does not clearly cover generative AI tools students use directly, such as chatbots that help with homework. A “Digital Omnibus” revision also moved the Annex III compliance deadline to December 2027.

Article 4 is already binding. Since February 2025, companies deploying AI have been required to ensure “sufficient AI literacy,” an approach similar to the literacy requirements New York is introducing.

New York Sets Broad Student Restrictions

New York City has become the largest US school district to adopt comparable limits. Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels announced a one-year moratorium on student-facing generative AI for children from 2-K through eighth grade.

The policy affects nearly 600,000 students, about two-thirds of public school enrollment. More than 38 approved classroom programs that fail new safety requirements will have their AI functions disabled, while companion chatbots will be banned across all grades.

Teachers may use AI for lesson planning but not grading or crisis management. Up to 50,000 high school students can join supervised pilots, limited to five classes per school, while other high school students will receive AI literacy lessons twice a year.

“The tech industry wants us to believe that AI-powered early education is not only inevitable, but necessary,” Mamdani said. “We do not see it that way.”

U.S. Schools Lack A Common Approach

Los Angeles Unified, the nation’s second-largest school district, is temporarily blocking generative AI tools on student devices across all grades while developing a permanent policy. It also banned school-issued devices for its youngest students in June as part of a broader screen-time strategy.

At least 37 US states have issued official AI guidance for schools, but there is no common definition of AI literacy or national agreement on how it should be taught.

Schools Still Face Questions

Michael Mulgrew, president of the United Federation of Teachers, welcomed New York’s screen-time limits but questioned how the Department of Education will verify that purchased software includes adequate safeguards.

Josh Golin, executive director of Fairplay, said a one-year review period is too short and argued that AI companies should demonstrate that their products are safe, effective and not designed to encourage cheating.

AI is already in classrooms, leaving schools and regulators to determine who sets the rules, who bears the risks and whether existing safeguards can protect children.

New York City public school students are scheduled to return to classrooms next week.