New York City Public Schools are giving teachers and students room to explore artificial intelligence while keeping important decisions in human hands. The citywide framework allows approved tools to support classroom learning and school operations, but bars AI from determining grades, discipline, academic placement or access to programs. The rules apply across grade levels and school settings, covering instruction, assessment, learner support and student services.
The policy uses a traffic-light system. Red uses include student promotion, graduation, disciplinary action, special education and 504 plans, counseling, crisis intervention and behavioral surveillance. AI-generated assessment information may inform an educator, but the teacher responsible for the course retains authority over what a student knows and the grade awarded.
Yellow uses require professional review and added safeguards. Students may use AI for research, exploration and creative projects when teachers provide guidance, evaluate the output and account for age and context. AI may also help identify patterns in school data, translate instructional material or prepare accommodations, although trained staff must review the results before they reach students.
Green uses cover lower-risk administrative and instructional support, including brainstorming lesson plans and drafting routine communications. Every tool that processes personally identifiable student information must pass the school system’s privacy and security review. Student data cannot be used to train AI models, sold or monetized, and unapproved tools cannot receive grades, identification numbers, special education status or other protected information.
The rules replace the patchwork that developed after the city briefly blocked ChatGPT in 2023 and individual schools began creating their own policies. Grade-specific guidance, academic integrity standards and a public inventory of approved tools remain under development. For families, the framework establishes a practical dividing line between AI as supervised classroom support and automated judgment over a child’s education.
