NYC bans student AI use through grade 8, sparks debate on tools

New York City has imposed a sweeping ban on student-facing generative AI for students from pre-K through eighth grade, with limited exceptions for AI literacy programs in high schools. The policy, announced by Schools Chancellor Kamar Samuels and Mayor Zohran Mamdani, permits teachers to use approved AI tools for planning and operations but prohibits their use for grading, IEPs, or behavior tracking.

The restrictions include carve-outs for multilingual learners and students with disabilities, according to details first reported by the New York Daily News.

The case against broad categorization

Abbas Manjee, a former South Bronx math teacher and co-founder of education technology company Kiddom, argues the ban fails to distinguish between fundamentally different technologies. Writing in the Daily News, Manjee contends that grouping all AI under a single label is like treating "a slot machine and a stethoscope" as the same thing.

The policy stems from draft guidance Samuels presented in March using a color-coded framework that Manjee says never addressed core questions about student data and cognitive development. By May, the chancellor acknowledged the city "missed the mark" and opted to ban the category rather than refine the criteria.

Manjee points to his company's experience in District 11, where nearly 6,000 middle school students and 180 teachers used Kiddom's scoring tool last year. Students whose teachers used the platform outperformed peers by up to 16 percent, and District 11 gained roughly 11 points on state tests compared to a similar district. In July, his team demonstrated to city officials that Kiddom met every requirement in the guidance and addressed parent privacy concerns, but received no response.

Why it matters

The tension reflects a broader challenge facing school systems nationwide: how to regulate emerging technology without blocking tools that demonstrably support teaching and learning. New York City has simultaneously asked teachers to adopt rigorous shared curricula through initiatives like NYC Reads and NYC Solves while potentially removing the software designed to help implement those programs. The result may preserve the mandate while eliminating the infrastructure teachers need to execute it outside classroom hours.

A functional alternative

Manjee proposes writing regulations at the level of function rather than label, distinguishing between student-facing chatbots and teacher-facing operational tools. He advocates requiring evidence as a prerequisite for any technology used with New York students, with companies obligated to publish results. Most would fail that standard, he notes, which is the point.

He also criticizes the city's communication approach, noting that Mayor Mamdani previously explained complex policy like a pied-à-terre tax through accessible three-minute videos but has not applied the same public education effort to explaining why different AI tools pose different risks and benefits.

The opinion piece was published in the New York Daily News.