Is ChatGPT safe for schools?
When a teacher types into a public AI chatbot, student data leaves the school and nobody checks the answers. Here are the real risks and the safer options.
Short answer
ChatGPT can be used safely in a school only with clear rules. The main risks are student data being sent to an outside company, and answers that sound right but are wrong. Personal accounts carry the most risk. School plans reduce it, and on-premise AI keeps the data inside.
What happens to the data a teacher types
When a teacher types into a public chatbot, the text is sent to the servers of the company that runs it. If the text has a student name, marks or a health note, that personal data has now left the school.
On a free or personal account, conversations may also be used to improve the model unless the user turns this off in the settings. The company says its business and education plans do not train on customer data by default. Terms change, so read the current terms before you decide.
The four main risks
These apply to any public AI chatbot, not only ChatGPT.
- Privacy: personal data about children is shared with a third party.
- Accuracy: the tool can state wrong facts with full confidence.
- Control: the school cannot see what staff typed or where it is stored.
- Law: data protection rules for children may require consent and care that a personal account does not give.
Age rules for students
Public chatbots set their own age limits. ChatGPT terms ask users to be at least 13, and users under 18 need permission from a parent or guardian. Your country may have stricter rules for the personal data of children. For this reason many schools let staff use AI but do not let students open their own accounts.
Three ways a school can use AI
Each step gives the school more control over its data.
- Personal accounts: fast and free, but the school has no control. Staff should never type student data.
- A school or education plan of a public tool: better terms and admin controls, but the data still sits in the vendor cloud.
- On-premise AI: the AI runs on servers the school controls, so prompts and answers stay inside the school.
Questions to ask before you approve a tool
Ask the vendor to answer these in writing.
- Where is our data stored, and in which country?
- Is our data used to train or improve your models?
- Who in your company can read our conversations?
- How do you check that answers are correct?
- Can we delete all our data, and how fast?
Frequently Asked Questions
Short answers to the questions education leaders ask most.
Yes, for general work with no personal data, such as drafting a lesson outline. They should not type student names, marks, behaviour notes or any other personal detail into a public chatbot.
On free and personal accounts, chats may be used to improve the model unless the user turns this off. The company says business and education plans do not train on customer data by default. Check the current terms.
A ban is hard to enforce and staff may keep using personal accounts quietly. A clear AI policy and one approved tool usually work better than a ban.
Keep student data inside systems the school controls, give staff an approved tool, and choose a tool that checks its answers. On-premise AI is the option that keeps the most control with the school.
Keep reading
More plain guides for leaders of schools and universities.