top of page
Search

I've Submitted Evidence to the Education Select Committee on AI in Schools


Parliament is asking the right questions about AI in education. This week, the Education Select Committee launched an inquiry into the use of artificial intelligence and EdTech across the education system, from early years through to universities. They want to know how AI is reshaping teaching, what the risks are, and whether the Government has an adequate framework to steer adoption responsibly.



I've submitted written evidence, drawing on my experience as both an assistant headteacher responsible for teaching and learning and as a co-founder of an AI coaching platform for teachers.


My submission focused on three areas.


The coaching bottleneck


We know from the research base what effective teaching looks like. The Great Teaching Toolkit Evidence Review makes that clear. The problem has never been knowing. It has been delivering feedback that is specific, timely, actionable and regular enough to make a difference. In most schools, the dominant model remains the formal lesson observation: infrequent, resource-intensive, and often perceived as evaluative rather than developmental. AI-powered transcript analysis offers a way to break that bottleneck, giving teachers access to detailed, evidence-grounded coaching reports without requiring an observer to be physically present.


The coaching vs compliance line


This is the critical design question for any AI tool that analyses teaching. If teachers perceive it as surveillance, adoption fails and trust erodes further. I argued that certain principles must be non-negotiable: teacher-initiated recording, private-by-default feedback, no compulsory scoring, anonymised leadership data, teacher control over deletion, and explicit separation from performance management. These are not nice-to-have features. They are the conditions under which the profession will engage.


The quality assurance gap


Schools are making purchasing decisions about AI tools with very little ability to evaluate whether those tools are educationally sound, ethically designed, or technically safe. I recommended that the Government develop clear quality standards for AI tools that analyse teaching practice, and that the DfE provide standardised data protection guidance to reduce the burden on individual schools.


I was also able to reference recent research from Stanford and elsewhere showing that automated feedback on classroom discourse can meaningfully improve teaching practice when designed well. The evidence base is growing, but the regulatory and quality assurance framework has not kept pace.


I am presenting peer-reviewed research on this topic at the BERA Teacher Education Advanced Network (TEAN) Conference at Sheffield Hallam University in May. If Parliament is serious about getting AI in education right, practitioner voices need to be central to that conversation.


The inquiry deadline for written submissions is 10 April 2026. If you are a school leader, teacher, or EdTech professional with experience of AI in the classroom, I would encourage you to contribute. Details here:



Adam Sturdee is a senior leader and co-founder of Starlight, the UK’s teacher-first AI-powered transcript-based coaching platform for educators. His work sits at the intersection of dialogic practice, instructional leadership and responsible AI strategy for schools and trusts.


He will be presenting his research on AI-supported coaching at the BERA TEAN Conference 2026: https://www.bera.ac.uk/conference/bera-tean-conference-2026


If you would like to explore these ideas further:


Learn more about Starlight: https://www.starlightmentor.com

Read more on AI and coaching: https://www.coaching.software

Enquire about speaking or consultancy: https://www.adamsturdee.com/consulting

 
 
 

Comments

bottom of page