Guide for school leaders
AI in schools and children's data: what to check before a pilot
A practical checklist for private school leaders in Qatar, written with Qatar's Personal Data Privacy Protection Law (Law No. 13 of 2016) in mind. Use it for any AI tool, not only Murshid.
Purpose and scope
- Write down what the AI will do, for which grades and subjects, and for how long.
- List the student data the tool needs, and leave out anything it does not need.
- Name the person at the school who owns the pilot and answers questions from families.
Children's data and guardian consent
- The PDPPL gives children's data special protection and requires permission before it is processed. Plan how guardians give that permission before any student uses the tool.
- Explain to families, in plain language and in their language, what the AI does and does not do.
- Check that a guardian can withdraw consent, and that withdrawing actually stops processing for that child.
Teachers stay in charge
- Confirm that a teacher decides what the AI can do for students, class by class.
- Check whether a student tutor hands over answers or guides the student to them.
- Make sure teachers can see what the AI says to their students.
Records you can show
- Ask whether every AI interaction is recorded, and who at the school can see the records.
- Ask whether you can export the records and consent evidence for your board or an inspection.
- Ask how changes to past records would be detected.
Questions for the vendor
- Is student data ever used to train AI models?
- Where is student data stored and processed, and by which providers?
- What happens to the data when the pilot ends?
- How, and how quickly, would the school be told about a data incident?
After the pilot
- Review the results with teachers and families before deciding to continue.
- Delete or return the data you no longer need.
Murshid
How Murshid answers these questions
Guardian consent that blocks the tutor when revoked, a teacher in the loop, a tamper-evident audit log you can export, and no training on student data.