Teachers and School AI Policy: Cambridge's 8 Principles

In early October, Cambridge University Press & Assessment published a report called Shaping AI Policy and Practice in English Language Teaching. It draws on teachers around the world who adopted generative AI early, along with earlier Cambridge research on how teachers use it. The main finding is simple: many teachers are working out how to use AI on their own, with little direction from the people writing policy.

This article is based on Cambridge's announcement and press coverage, not the full report text.

Who the report is about

The press release talks about schools in general, but the research behind it comes from English language teaching. That matters when you read the recommendations. They are a good starting point for any subject, but they were not tested across every classroom.

The eight principles

Cambridge groups its advice into eight points:

Teacher and learner input. Teachers and students should help decide how AI tools are chosen, introduced and judged.
Ongoing training. Teachers need continuing support, not a single workshop.
AI literacy. Students should learn to question AI answers, spot errors and bias, and know when not to use it.
Assessment that shows the process. Grading should value drafts, reflection and independent thinking, not only the finished piece.
Human skills first. Policy should aim at critical thinking, communication and social responsibility.
Trust and safeguards. Clear rules on privacy, age-appropriate access, transparency and accountability.
Professional authority. Picking, adapting and reviewing AI tools should sit with the teachers who use them, not arrive as a finished decision from above.
Fair access. Tools should work across languages and in under-resourced schools so gaps don't widen.

Turning the principles into a buying test

The report reviews no individual products. What follows is our own reading of it. Take any classroom AI product and hold it up against three of the principles.

Start with teacher agency. Ask the vendor which teachers used the product before launch and what changed because of their feedback. A tool shaped by classroom input usually has a story about something that was removed or reworked.

Next, look at assessment. A product that only hands students a finished answer works against the process-focused approach the report describes. One that keeps drafts, prompts and revision history gives a teacher something to mark besides the final text.

Last, test fairness and safeguards together. Open it in a language other than English, on an older laptop, with a student account set to the youngest age you teach. Then read what the privacy page says happens to student writing. Whatever breaks first will tell you more than a sales demo.

What the report doesn't tell you

The principles are recommendations, not measured results. The report doesn't show that schools following them get better outcomes, and it doesn't compare specific tools. Treat it as a checklist for asking better questions, not a certificate of quality.

It also reflects one subject area. A maths teacher or a university lecturer may weigh some points differently.

Where to start

If you are a teacher, ask your school who currently decides which AI tools are approved, and whether you can join that conversation. If you are a school leader, the first and fourth principles are the cheapest to act on: put teachers in the room when tools are chosen, and look at how your assessments treat process.

To compare tools built for classrooms, see our AI for Education directory.

Update log

October 2026: First version, based on Cambridge's announcement and press coverage. Next review: November 2026.

Sources: