Should schools ban AI?
A blanket ban rarely works. Students already use AI outside school, and New York City reversed its 2023 ChatGPT ban within months. Most schools do better with clear rules: age-appropriate tools, assignments that show students' thinking, limits on AI in assessed work, and teaching students when AI helps and when it harms learning.
Most schools should not ban AI outright, but they should restrict it. A blanket ban is hard to enforce, does not stop use at home and leaves students without guidance. Unrestricted use is not the answer either. Research shows that unguided AI can hurt learning. The workable middle is clear rules about which tools, for which tasks, at which ages.
What happened when schools tried a ban
New York City's Department of Education blocked ChatGPT on school devices and networks in January 2023. By May, Chancellor David Banks reversed the ban, describing the first response as "knee-jerk fear" that overlooked the technology's potential, and promised teachers resources and examples instead.
Meanwhile, students kept using AI. Common Sense Media's 2024 survey found that 7 in 10 US teens had used at least one generative AI tool, and 83% of parents said their child's school had not communicated with families about generative AI.
What the research says about AI and learning
The best-known randomized study so far, published in PNAS in 2025, gave nearly 1,000 high school math students in Turkey access to GPT-4:
- Students with a plain ChatGPT-style tool did much better during practice, but scored 17% worse than the control group once the AI was taken away.
- Students with a tutor version designed to give hints rather than answers did far better in practice and largely avoided that drop.
The lesson for school leaders: the question is less "AI or no AI" and more "which AI, used how". Tools that hand over answers can become a crutch. Tools that make students do the thinking can help.
Arguments on each side
| For restricting heavily | For allowing with rules |
|---|---|
| Adult chatbots are not designed for children and their terms usually require age 13+. | Students will meet AI anyway and need guidance to use it well. |
| Unguided use can replace practice students need. | Guided, hint-based tools can support learning. |
| Assessment integrity is harder to protect. | Teachers can redesign tasks to show process. |
| Privacy risk if student data goes into consumer tools. | Approved, school-contracted tools keep data under school control. |
A middle path for K-8 schools
- No personal accounts on adult chatbots for under-13s. Teachers can demonstrate them on the board.
- Approve a short list of tools built for children, with school data agreements and teacher visibility.
- Define AI-free zones: tests, spelling and handwriting practice, and core skill practice.
- Define AI-welcome zones: brainstorming, asking follow-up questions, explaining a concept another way.
- Teach AI literacy so students understand that AI can be wrong.
- Tell parents which tools are used and why.
Templates help: see AI policy for schools: a free template and How schools can pilot AI.
If you want a classroom tool that fits these rules, Askie for Schools is built for ages 4 to 15, gives each student a private access code and lets teachers review conversation history, and its free tier covers 20 students.
Bring safe AI into your classroom
Askie for Schools brings a safe, age-appropriate AI learning companion to your students, managed from your teacher account.