Getting AI in education right: reflections from the Cambridge Summit

AI is already changing education, but how schools and colleges respond to it matters just as much as the technology itself. While AI has the potential to improve access to learning and reduce teacher workload, it also raises important questions about safeguarding, privacy, professional judgement and the role of teachers.

One thing became clear at the Cambridge AI in Education Summit 2026: nobody has all the answers yet. For schools, this makes it even more important to develop an informed, responsible approach to AI - one that puts education, people and learning first.

Earlier this month, AiEd Certified Programme Director, Dr Sarah Alcock attended the summit, hosted by Cambridge University Press & Assessment, where she presented early evidence from AiEd Certified in the Responsible AI implementation strand. Bringing together researchers, teachers and policymakers, the event explored some of the key challenges and opportunities facing education. Here, Sarah shares three themes that stood out to her and what they mean for schools and colleges.

This month I spent two days at the Cambridge AI in Education Summit 2026, hosted by Cambridge University Press & Assessment. I was there to present early evidence from AiEd Certified in the Responsible AI implementation strand (abstract). Sitting among researchers, teachers, and policymakers, I was reminded how much perspective a room like that gives you.

Over the two days, three ideas kept coming back. Each one says something about why we do what we do at AI in Education.

1. Learners need rights, privacy and human judgement

The opening session raised an immediate and wide-ranging problem: giving every learner the same tool says nothing about whether they can think well with it. It came from Professor Ann Kristin Glenster (LOGOS, University of Cambridge) and Megan Ennion (DEFI), who are working towards a Learner Epistemic Rights Charter.

Learners need privacy to make mistakes as they’re working. A student who has put everything into an essay might also feel disappointed to have it marked by AI, a concern that came up again in a later panel. Many schools are trialling AI marking, for very good reasons: workload is real, and teacher time is precious. I don’t think the answer is to stop, but I would invite every school to reflect on what their learners understand about this process. Do your students know when AI is involved in marking their work, and why? Do they know where a teacher still reads, responds and decides? Being open with learners about the practical and administrative reasons, and keeping the human connection visible, helps them understand a decision they might otherwise experience as unfair.

These questions and decisions should be defined in an AI policy, and reflected in how your school or college listens to and talks with its students.

2. Teachers are under pressure to adopt AI

Several speakers described teachers being pushed into using AI. Jaione Santurtún (CALL) drew a parallel with earlier waves such as gamification. Dr Ricardo Lozano (Texas A&M International University) reminded us that teachers’ beliefs, values, identities and loyalties shape what happens in any classroom. Dr Kristina Kallas, Minister of Education and Research in Estonia, described how Estonian teachers meet regularly in ‘learning circles’ to learn together and compare AI practice.

This is the heart of why AiEd Certified exists. We support schools and colleges to adopt AI meaningfully, as a whole-school approach led by education rather than technology. Handing staff a new tool and hoping for the best rarely works. Collaboration & Community is one of the five elements of the framework for exactly this reason. It asks leaders, staff, students and families to learn about AI together and share what works. At Explorer level that may be informal conversations; at Practitioner level, coordinated training and subject working groups; at Innovator level, sharing practice with research bodies and beyond. Estonia’s learning circles are a national-scale version of the same idea.

3. “Nobody has this sorted”

In the Collective Intelligence panel, Professor Rose Luckin said: “Nobody has this sorted, even the ones who look confident.”

I found that reassuring. Dr Kevin Martin (DEFI) called for safe spaces for design-based research, and the panel argued for collective evidence and judgement. That fits how I think about certification. Being certified doesn’t mean a school has AI sorted. Confidence with AI is about being able to ask good questions, review your practice and gather evidence of what works in your own context. The framework gives schools a structure for doing that, and a community to do it with.

What does this mean for schools?

The summit reinforced that responsible AI adoption isn't about having all the answers from the outset. It's about creating the right conditions for schools and colleges to keep learning, ask questions and make informed decisions as the technology develops.

AiEd Certified provides a framework to help education settings do just that, bringing together AI Literacy, Policies & Ethics, Tools & Systems, Digital Pedagogy, and Collaboration & Community.

Find out how AiEd Certified works or book a discovery call to find out more.

Dr Sarah Alcock, Programme Director, AI in Education

‍

Header image credit: Cambridge University Press & Assessment