Policymakers must give teachers a stronger voice on AI policy in schools, new Cambridge report recommends

New Cambridge University Press & Assessment report sets out eight key principles for effective AI integration in schools

Policymakers should work closely with teachers to shape policy for the AI age in schools, and invest in comprehensive, sustained teacher professional learning to ensure educators can effectively use AI in their classrooms and critically engage with tools, a new Cambridge University Press & Assessment (Cambridge) report recommends.

Crucially, teachers and learners should contribute to the design, procurement, implementation and evaluation of GenAI tools so that policy is grounded in classroom realities.

The new report, ‘Shaping AI Policy and Practice in English Language Teaching’, draws on the best practices of English language teachers from around the world who have been early adopters of GenAI, as well as earlier Cambridge research on AI and teacher use, seeking to bring structure to an often-disjointed approach to AI use and adoption.

The findings suggest many teachers globally are navigating AI, but in a vacuum, often without clear direction. Teachers’ number one concern is how to support students to become AI-ready – to critically analyse AI outputs, proactively question their accuracy, and skilfully use GenAI as a tool to support, rather than shortcut, learning.

To support teachers and inform policy, Cambridge sets out eight key principles for harnessing AI effectively in schools:

• Involve teachers and learners in AI governance. Teachers and learners should contribute to decisions about the implementation and evaluation of GenAI tools so that policy is grounded in classroom realities.
• Invest in sustained teacher professional learning. Teachers need ongoing support to evaluate tools critically, make informed pedagogical decisions, adapt GenAI to learning goals, and teach students responsible use.
• Embed critical AI literacy as a curriculum expectation. Students should learn how to question AI outputs, identify bias and inaccuracies, use GenAI as a learning support rather than a shortcut, and understand when not to use it. Critical AI literacy is essential and not an optional add-on.
• Reform curriculum and assessment frameworks. Curriculum and assessment systems should recognise process evidence, feedback literacy, learner reflection, and independent thinking alongside final products, in an integrated learning and assessment approach.
• Prioritise human-centred educational outcomes. Policy frameworks should focus on the human capabilities that need to be strengthened in the AI era, such as critical thinking, communication, autonomy, emotional intelligence and social responsibility.
• Build trust through ethical use and robust safeguards. Responsible integration and trust by all stakeholders, e.g., teachers, learners, parents, depends on clear approaches to data privacy, age-appropriate access, inclusivity, responsible authorship, transparency and accountability.
• Protect teacher agency and professional judgement. Policies should position teachers as decision-makers in tool design selection, classroom integration, evaluation and improvement – not merely as implementers of externally chosen technologies.
• Ensure equitable access to high-quality tools. Policy should address language diversity, local contexts, accessibility, and under-resourced settings so that GenAI does not widen existing inequalities.

Dr Evelina Galaczi, Director of Research, English, at Cambridge University Press & Assessment, said:

“Policymaking has often struggled to keep pace with the rapid rate of technological change, but teachers have had to adapt. This has led to gulfs in knowledge, and an inconsistent approach to AI education and adoption. As often happens, it has become the quiet responsibility of teachers to develop the practical knowledge of what responsible AI integration actually looks like.

“Teachers who are already successfully experimenting with AI in their classrooms are prioritising professional judgement, human-centred practices and critical AI literacy. By following this lead, we can ensure AI becomes an effective and responsibly applied tool for teaching and learning, not a crutch for learning shortcuts.”