The Relationship Questions We Should All Be Asking

I want to talk about relationships.

I know – from my British point of view, this sounds like the beginning of an episode of a slightly dodgy sitcom. However: many across education are saying “we need to get AI to do things so that people can be freed up to do the things that humans do best”. Unfortunately, I don’t think it’s that simple, for three key reasons.

First, we do have a lot of evidence that relationships between teachers, students and parents matter for educational outcomes. However, we don’t yet have much evidence on what aspects of the educational process AIs can and should take over, assuming that we are seeking some combination of better learning outcomes and improved student wellbeing (note that I am talking about AIs, plural: there is no single, canonical technology here). And we don’t know which tasks and roles for AIs change the relationships between teachers, students and parents, or how. There is already some controversy here: for example, I have heard reports of people arguing that teachers should continue to do marking of homework which is often seen as tedious, because it is a good way of their really understanding the mistakes students are making; but others are vociferous about how AI marking can and should save teachers’ time.

Secondly, AIs are already changing relationships. Students are already using conversational AIs for homework, and teachers rarely know much about these transactions. Worse, conversational AIs – particularly generic ones not designed for learning – are frequently designed to be sycophantic and offer an attractive, engaging conversation. Sometimes this might be more appealing, if not more worthwhile, than a sensibly demanding teacher. And worst of all, some AIs are (in my personal view) fundamentally unsafe, and may lead to young people having damaging understandings about what it means to have relationships with human beings: the work of Michelle Culver and team at the Rithm Project, Julia Freeland Fisher and the EdSAFE AI Alliance, led by Erin Mote, are all well worth engaging with on this topic.

Thirdly, relationships have never been more important. Of course, understanding how to build healthy and meaningful connections with other human beings is essential in itself, but it’s now become a vital skill to thrive in the working world, too (see the recent work of the National Foundation for Education Research in the UK). “Relational Education” – a phrase I first heard from Isabelle Hau, who continues to write compellingly about it – has never been more important.

So, let’s talk about relationships: how AI can enhance rather than harm them in education, how we teach them better, what good relational education looks like, and what research and actions we now need as education players concerned with impact. And, of course – I look forward to discussing all of this with you as part of our relationship in 2026!