Give Teachers Voice in AI Policy in Schools, Cambridge Report Urges

The call for a stronger teachers voice in AI policy in schools now has a detailed blueprint behind it. On 4 November 2026, Cambridge University Press & Assessment published a report arguing that policymakers should work closely with teachers to shape rules for the AI age, and that teachers and learners should help design, procure, bring in and evaluate generative AI tools so policy stays grounded in classroom realities (Cambridge via PR Newswire). Information verified through 7 November 2026.

The report, Shaping AI Policy and Practice in English Language Teaching, draws on English language teachers around the world who adopted generative AI early, plus earlier Cambridge research on AI and teacher use (Cambridge via PR Newswire). Its underlying interviews covered 15 teachers and teacher trainers in 13 countries across six continents, interviewed in early 2025, with experience ranging from five years to more than twenty (WINS Solutions).

That is a small, self-selected group. Cambridge itself notes the findings reflect early-adopter practice, not a representative survey of all teachers, and the report does not measure student outcomes at scale (WINS Solutions). Read it as experienced practitioners describing what worked, not as proof of what works everywhere.

Teachers voice in AI policy in schools: what the eight principles actually ask for

Cambridge sets out eight principles. Teachers and learners sit inside AI governance. Schools invest in sustained teacher professional learning for AI. Critical AI literacy becomes a curriculum expectation, not an optional extra. Curriculum and assessment change to value process, feedback literacy, reflection and independent thinking alongside final products (Cambridge via PR Newswire).

The remaining four push in the same direction. Policy should prioritise human-centred outcomes such as critical thinking, communication, autonomy, emotional intelligence and social responsibility. Trust needs ethical use and safeguards on privacy, age-appropriate access, inclusivity and authorship. Teacher agency and professional judgement get protected — teachers as decision-makers, not implementers of tools chosen elsewhere. And access to good tools must be equitable, accounting for language diversity, local context and under-resourced settings (Cambridge via PR Newswire).

None of this is abstract in the teachers’ own accounts. They reported using generative AI for lesson planning, adapting materials to different proficiency levels, generating quizzes, giving feedback and marking, and producing classroom resources (WINS Solutions). Their worries were practical too: accuracy, bias, privacy, and students leaning on AI before attempting a task themselves (WINS Solutions).

The number one concern: critical AI literacy as a curriculum expectation

Cambridge found that teachers’ top concern is how to make students AI-ready — able to analyse AI outputs critically, question their accuracy, and use generative AI to support learning rather than shortcut it (Cambridge via PR Newswire). That is why the report insists critical AI literacy as a curriculum expectation belongs in the framework itself. Students should learn to spot bias and inaccuracies, and to know when not to use a tool at all (Cambridge via PR Newswire).

Dr Evelina Galaczi, Cambridge’s Director of Research, English, put the gap plainly. “Policymaking has often struggled to keep pace with the rapid rate of technological change, but teachers have had to adapt,” she said, adding that responsible AI integration had become “the quiet responsibility of teachers” (Cambridge via PR Newswire). Teachers in the study overwhelmingly rejected the idea that AI could replace them; one Australian teacher quoted in the report said educators remain “the experts in the classroom” (WINS Solutions).

A generative AI use in higher education perception gap points the same way

Universities are having a parallel argument, and the numbers are awkward. Research from the University of Manchester and the Universidad del País Vasco, published in Frontiers in Education, surveyed 45 academic staff and 86 students at a UK research-intensive university between April and June 2024 (Phys.org). Asked how much students use AI for academic work, staff gave an average rating of 5.23 on a seven-point scale. Students rated their own use at 3.11 (Phys.org).

Who is right? The researchers caution that students may have underreported, but the gap itself shows how little reliable information staff have about actual practice, since they usually see only a final assignment, not the process behind it (Phys.org). Students were also more positive than staff about AI’s benefits for learning, productivity and workload, and reported using it to find answers, generate written work and improve grammar or translation (Phys.org).

Staff familiarity varied widely. While 63% had used at least one generative AI tool in the previous six months, only 14% had a paid subscription (Phys.org). The authors argue the problem cannot be fixed with technical training or student AI literacy alone, and warn that assuming students use AI mainly to circumvent assessment risks building a relationship based on distrust (Phys.org).

Sustained teacher professional learning for AI, not one-off training days

Both studies land in the same place. Rules written without the people who live under them rest on guesses. Cambridge’s answer is sustained teacher professional learning for AI — ongoing support so teachers can evaluate tools critically, make informed pedagogical decisions and teach responsible use (Cambridge via PR Newswire). The Manchester team similarly calls for teaching and assessment to evolve, with staff and students supported to explore appropriate uses together (Phys.org).

Analysis: the strongest claim in the Cambridge report is also the most testable. If teachers help choose the tools, the tools should fit real lessons better, and the rules should be ones classrooms can actually follow. What neither document can yet show is whether that involvement raises student outcomes — the Cambridge work never measured them at scale, and the Manchester survey is a snapshot of one institution in 2024. The case for a teacher voice is convincing on governance grounds. The learning evidence is still to come.

Photo: Harrison Keely, via Wikimedia Commons (CC BY 4.0)

Sources

1. https://en.prnasia.com/story/554715-0.shtml
2. https://www.winssolutions.org/cambridge-ai-english-language-teaching-report/
3. https://phys.org/news/2026-11-students-ai-differently.html