New York City’s Department of Education (NYC DOE) released preliminary guidelines for AI use in schools on March 24, 2026 — the first formal citywide AI guidance, arriving nearly three years after a short-lived ChatGPT ban.
Framework: The guidance uses a “traffic light” model:
- Green (approved): AI for lesson brainstorming, drafting communications, scheduling, resource planning, teacher training.
- Yellow (conditional, requires human review): Identifying trends in student data, generating translations for bilingual learners, adapting materials for students with disabilities.
- Red (prohibited): Grading, discipline decisions, development of Individualized Education Programs (IEPs) or 504 plans, counselling and crisis intervention, academic placement decisions.
Key principles: All AI tools must operate under human oversight; AI supports but never replaces educator decision-making. Student personally identifiable information (PII) may not be entered into any AI tool that has not completed the city’s ERMA (Enterprise Request Management Application) vetting process. Student information cannot be used to train AI models or generate revenue.
Limitations: The guidance does not establish an AI literacy curriculum or require students to learn AI. It governs educator and institutional AI use only. Issues including student homework use, personal AI accounts, algorithmic bias review, and instructional effectiveness standards remain unresolved and are expected to be addressed in the full June 2026 playbook. The ERMA vetting process currently evaluates data privacy and security but not algorithmic bias or instructional effectiveness.
Public feedback was sought through May 8, 2026.
Score note: No score change. The guidelines are a framework for responsible AI use by educators, not a curriculum mandate or AI literacy requirement. None of the scoring dimensions are directly met by guidance of this type.