New York prohibits AI in primary schools, three months following Norway's lead.

      New York City is implementing a one-year moratorium on generative AI for students from pre-kindergarten through eighth grade, prohibiting the use of AI software and companion chatbots for approximately 600,000 children. In June, Norway instituted a ban on generative AI in its primary schools, and the EU AI Act regulates AI utilized on students instead of AI used by them. Starting with the 2026-27 school year, New York City's prohibition will apply to students from pre-school to eighth grade, according to ABC News.

      Mayor Zohran Mamdani characterized this decision as a rejection of the narrative that A.I.-powered early education is both inevitable and essential. “We do not share that perspective,” he stated. The district is opting to eliminate the software entirely rather than providing guidance, banning all generative AI accessible to students along with companion chatbots.

      High schools are mostly exempt from this ban, with up to 50,000 students permitted to participate in five pilot programs, limited to five classes per school, alongside semi-annual literacy modules for other students. This is not the first instance of such a ban, as Norway prohibited generative AI for grades one to seven in June, allowing supervised use up to age 16. That ban came into effect at the end of August, with Prime Minister Jonas Gahr Store noting the increased risk of students bypassing essential educational steps due to AI.

      While Europe has already enacted these bans, it lacks a comprehensive AI rulebook addressing these issues. The AI Act categorizes educational applications as high risk under Annex III, which includes aspects like admission decisions, learning outcomes evaluation, appropriate education level assessment, and monitoring during exams. Each of these refers to AI utilized on students by institutions, covering systems that admit children and detect cheating during tests, rather than addressing chatbots that assist with essay writing.

      The requirements outlined in this rulebook are not yet in effect; the Digital Omnibus moved the Annex III deadline to December 2, 2027. However, one member state, Italy, has taken action: since October of last year, Law 132/2025 mandates parental consent for the use of AI by those under 14.

      At present, the only EU-wide obligation in effect pertains to literacy. Article 4 compels providers to ensure adequate AI literacy since February 2025, which aligns with New York’s requirement for high school literacy modules. Thus, a similar trend continues, with restrictions on children and AI emerging from education ministries and national parliaments, as seen when Italy set a concerning example in 2023, rather than initiatives stemming from Brussels.