New York City's sudden restriction on classroom artificial intelligence has left schools scrambling and edtech companies defending their products. Just days before the school year began, the city disabled student-facing AI features in 38 software tools. This sudden move left teachers and parents confused about which programs are actually allowed in classrooms. Parents and educators must now figure out the murky boundaries between helpful learning technologies and banned digital tools.
What Happened
In September 2026, Mayor Zohran Mamdani announced a one-year moratorium on student-facing generative AI for grades 2-K through 8th in New York City public schools. The policy aims to protect student development and enforce new screen-time limits.
The city flagged 38 educational products, but kept the list private until a Chalkbeat investigation revealed the affected tools. Among those banned outright was Amira, an AI-powered reading tutor used in over 200 city schools, and the Education Department's own Microsoft-partnered chatbot.
Other companies are pushing back. HMH, which provides the city’s mandated reading programs like Read180 and Arriba la Lectura, insists its products do not use the prohibited student-facing generative AI. Translation tool Wordly was also flagged, despite the company's claims that it does not use generative AI and should qualify under the city's carveouts for English language learners.
The Bigger Picture
The developer of Amira claims its software is built on the Science of Reading and is backed by more than 100 studies showing reading gains. However, in school districts like San Francisco, parents and teachers have reported severe frustration because the speech-recognition AI frequently fails to understand children's voices.
To avoid these technical hurdles, some schools are turning to human-led virtual tutoring. For instance, the 'Air Reading' program, which uses live, certified teachers rather than AI, was shown in a randomized evaluation by Accelerate to generate an 11-percentile-point gain for struggling readers, though it costs schools $1,500 per student.
Meanwhile, other districts continue to use generative AI while heavily vetting for student privacy. Systems like Allentown School District and Firstline Schools actively use platforms like Coursemojo and Google's Gemini enterprise version, but they mandate strict terms to ensure student data is not used to train external AI models. As we previously covered, other platforms are also adding strict teacher controls to manage AI interactions.
What This Means for Families
Students in NYC public schools will experience immediate changes in how they read and practice math. With student-facing AI disabled on dozens of apps, teachers must pivot back to traditional, human-led methods or non-AI software.
The restriction of tools like Read180 and Amira means struggling readers may see shifts in their daily intervention schedules. However, this shift aligns with growing parent concerns regarding screen-time overload. To help manage this balance, some educators are looking to blended solutions that bridge physical print and digital screens to keep kids engaged without over-relying on algorithms.
The policy also forces a pause to evaluate data privacy. Because many tech companies train their algorithms on user inputs, restricting generative AI protects children’s voices and personal data from corporate harvesting.
What You Can Do
- Ask your child's teacher which specific learning apps are currently used in their classroom and if any AI features have been disabled.
- Request human-led alternatives like high-dosage tutoring programs if your child is struggling with reading, rather than relying solely on automated educational software.
- Check data privacy agreements for any educational apps your child uses at home to ensure their personal data is not being used to train generative AI models.