AI in Education: Khanmigo, ChatGPT Edu, Homework Apps — 2026-10-29

Top developments

NYC Expands AI Ban in Public Schools

New York City’s Department of Education has released a list of banned AI products for its 1 million students, following a September pause on generative AI. The ban includes popular reading and math programs, student assessment tools, and even a chatbot developed by the DOE itself. This move reflects growing local resistance to unvetted AI tools in K-12 education.

Stanford Professor Ends Graded Homework Over AI Cheating

A Fields Medal-winning professor at UC Berkeley (often cited in similar academic integrity debates) has ended graded homework, stating that in-person exams are the only remaining solution to widespread AI use. The decision highlights the failure of current AI detection tools to distinguish between human and AI-generated work, forcing a return to traditional assessment methods.

China’s AI Learning Device Market Intensifies

Chinese media reported on the fierce competition among AI learning device manufacturers, including Zuoyebang (Gauth), Xiaoyuan, and iFlytek. With over 42 million primary and secondary school students in China, parents are increasingly relying on these devices for homework help and personalized learning. The market is seeing a surge in devices that integrate large language models to provide step-by-step solutions and automated grading.

UNESCO Issues Guidelines on AI-Enabled Violence

UNESCO released a new brief outlining seven actions to prevent AI-enabled violence in education, emphasizing safeguards against technology-facilitated harm. This aligns with broader global efforts to define ethical boundaries for AI in schools, as seen in Vietnam’s strategic objective to teach children AI proficiency and protection skills.

Local view

Sina Education (China) reported on the ethical dilemma of using large language models for homework assistance. The article, "Big Model Homework Tutoring? Children’s Use of Tools Must Have Limits," highlights the tension between instant AI-generated answers and the development of student writing and mathematical reasoning skills. It calls for stricter parental and educational oversight to ensure AI tools aid rather than replace critical thinking.

Context & numbers

• Student AI Usage: 95% of UK undergraduates used AI in 2026, with 12% submitting AI-generated text directly into assessed work.
• Detection Accuracy: Turnitin claims a sub-1% false positive rate for documents with >20% AI writing, but studies show false-negative rates of 10-30% on lightly edited AI text.
• Chinese Market Scale: Over 42 million primary and secondary school student families in China are currently evaluating AI learning devices.

On the radar

• Vietnam’s AI Curriculum: Vietnam is accelerating its plan to integrate AI proficiency into general education, aiming to protect children from AI risks while enhancing their digital literacy.
• Turnitin Policy Shifts: More universities are canceling or restricting AI detectors due to false positives, with Washington State University having already canceled its detector in 2026.