Navigating AI Friction and Future Work Anxiety in Schools with Dr. Iro Konstantinou

In this episode live from the Festival of Education at Wellington College 2026, host Daniel Emmerson speaks with Grainne Hallahan, Head of Community at Teacher Tapp. Grainne shares unique data-driven insights into how UK teachers feel about artificial intelligence, the shifting patterns in subject-specific AI adoption, and why asking precise, time-bound questions is essential for capturing classroom realities. She also offers practical guidance for school leaders on policy formulation, teacher anonymity, and effective survey design.

Episode Overview In this episode from the Festival of Education at Wellington College 2026, host Daniel Emmerson speaks with Victoria Hedlund, founder of GenEd Labs. Victoria discusses bias in generative artificial intelligence, the importance of inclusion, and what meaningful human oversight should look like in education. She also explores how educators can challenge AI outputs, balance productivity with equity, and preserve the essential human element in teaching. Key Topics & Timestamps 00:00 Putting Output on Trial 00:33 Introduction to Victoria Hedlund 01:37 Researching Bias and Human Oversight 03:11 Why Schools Must Focus on Bias 05:56 Sparks, Lightbulbs, and Gender Stereotypes 08:12 Frequency of Exposure: Teachers vs AI 09:09 Visual Representation and Image Bias 13:06 Questioning Assumptions and Model Iteration 17:09 Equity Versus Productivity 20:49 The Human in the Loop 24:30 Embracing Messy Humanness in Education

Foundational Impact is a Good Future Foundation podcast exploring how artificial intelligence is changing education. Through conversations with people whose work is shaping the direction education takes next, we examine the evidence, the challenges facing schools and the decisions being made across the sector. Hosted by Daniel Emmerson. Follow Foundational Impact on Spotify, Apple Podcasts or wherever you listen. Website: www.goodfuture.foundation

Episode Overview In this episode from the Festival of Education at Wellington College 2026, host Daniel Emmerson speaks with Daisy Christodoulou, Director of Education at No More Marking. Daisy discusses the role of artificial intelligence in educational assessment, emphasising the importance of actionable feedback over lengthy written notes, the mechanics of comparative judgment, and why handwritten work remains vital for learning. She also addresses the potential pitfalls of AI in schools and offers key guidance for school leaders navigating integration. Key Topics & Timestamps 00:00 Speed As Quality 00:36 Meet Daisy Christodoulou 01:57 Is Written Feedback Useful 02:29 Thermostat Feedback Model 03:32 Comparative Judgment Explained 04:53 AI Judges And Benefits 06:09 Avoiding The Doom Loop 10:48 Student Trust And Emotions 12:48 Fast Feedback Tradeoffs 14:59 Handwriting Versus Typing 17:14 Digitizing Handwritten Work 17:57 Assessment Tradeoffs And Standards 20:21 Advice For School Leaders 21:56 Hopes For AI In Schools

Our guest in this episode is Dr Fiona Aubrey-Smith, an independent researcher who studies the relationship between pedagogy and digital tools, and the lived experience of students in the classroom. Fiona opens by explaining why identical classroom practices can produce completely different outcomes depending on the belief system behind them. Her example of two teachers greeting students at the classroom door in seemingly the same way, but for very different underlying reasons, shows how policies alone don't determine impact; the intent behind them does. The conversation moves into belonging and dignity, including research showing that struggling students typically ask for help only 1-3 times in a lesson regardless of how much support they need, while confident students almost never ask to be pushed further. Daniel and Fiona discuss the tension between protecting a student's dignity and the value of productive vulnerability. From there, Fiona lays out her framework for introducing technology into schools: start with the purpose of school, define the role of the learner, describe what the learning experience should feel like, define the role of the teacher, and only then decide which tools will help deliver it. Starting with the tool itself, before asking why, is a costly and common mistake. Two school case studies illustrate this in practice. At Haileybury, an embedded one-to-one, on-demand digital provision (available at the point of need rather than simply issued to every student) has freed teachers to rethink classroom layout entirely. At Connect Education Trust, a similar model has extended digital skills and connectivity into the home, empowering families to support learning without needing their own digital capability, coinciding with outstanding inspection outcomes. On generative AI, Fiona shares some thoughtful use cases she sees, such as chaining AI tools to help secondary students revise, and using AI image generation to help primary students with their descriptive writing. She wraps up with a challenge for school leaders. Rather than starting with policies or tools, ask what your school would leave behind if it closed tomorrow, since that gap is where the real human value lies. From there, let AI and digital tools take on the automation and acceleration, so teachers are freed up for the parts of the job only a human can do.