AI could transform teacher development. But should it?

Most debates about AI relate to issues such as whether pupils are using ChatGPT to write coursework or how teachers are using AI to create lesson plans and prepare resources.

There’s no surprise that these are the main points of discussion - but there is another debate that is less visible but no less consequential: what happens when AI is used to develop teachers themselves?

It is a question worth asking because the possibilities are genuinely exciting. One of the biggest constraints on teacher development is time for mentoring, feedback and reflection.

How AI could support teacher development

AI could help with all of these. For example, trainee teachers could receive richer, more targeted feedback if AI was used to help mentors identify patterns in classroom practice that might otherwise have been missed.

It could also be deployed to help a school leader rehearse a difficult conversation with a staff member in a simulated environment, or even help a training provider identify early signs that a trainee is struggling, and put support in place.

Versions of this kind of support already exist, and some are beginning to shape teacher education and professional development.

These tools aren’t designed to replace mentors, coaches or teacher educators, but instead help existing ones use their time and judgement more effectively. For a profession facing persistent workload pressures, that could make a real difference.

The importance of mentoring

Mentoring illustrates this particularly well. High-quality, frequent feedback is one of the strongest drivers of professional growth, yet mentors often have limited time and competing demands.

Used thoughtfully, AI could help mentors to prepare for coaching conversations, identify patterns across observations and streamline routine tasks.

This would leave more time for the work that only humans can do, such as making nuanced judgements, understanding context and building the relationships that make mentoring effective.

The same is true of AI-powered simulations. Teaching and school leadership involve countless high-stakes moments that are difficult to practise safely, including managing challenging behaviour, navigating sensitive parental conversations or responding to safeguarding concerns.

Simulated environments can create opportunities to rehearse these situations before they occur in practice, and help to build teachers’ confidence and self-efficacy.

Yet this enthusiasm also carries a risk.

What AI can’t replace

In teacher development, what matters most is not always the visible activity but the thinking and reflection it prompts. Feedback, for example, is not just about giving comments.

Its real value lies in the learning it stimulates, including reflection, discussion and the gradual development of professional judgement.

That learning is often shaped through relationships with experienced colleagues who challenge assumptions, build confidence and help teachers to make sense of complex situations.

AI may become increasingly capable of reproducing the visible features of effective professional development. It can already generate feedback that resembles that of an experienced mentor.

What we know far less about is whether AI supports the deeper processes through which teachers develop expertise - professional judgement, reflective thinking and trusting relationships.

Those human aspects should remain central to teacher development, regardless of how sophisticated AI becomes.

For, just as research conducted in England found that pupils thought they would be less motivated if they believed their work was being marked by AI rather than by a teacher, adults may respond similarly.

If AI changes the way teachers engage with mentors, colleagues and professional learning, the consequences could be far more significant than any time or cost savings.

Teacher benefits must come first

Therefore, with any potential use of AI for teacher development, the key question must always be: will this help teachers to learn more effectively, develop stronger professional judgement and ultimately improve outcomes for children?”

This approach should also come with the recognition that many AI tools will be designed by a competitive, and as yet unregulated, industry.

That is why we need stronger evidence before AI becomes embedded across teacher professional learning.

We need clear principles for its use, training that helps teachers to engage with it critically, and independent research into its impact on teacher development, as well as pupil outcomes.

Ultimately, the goal is not more AI. The goal is better-supported teachers, making better professional decisions and achieving better outcomes for children.

AI may become part of the future of teacher education, but that future should not be defined by the technology itself.

It should be defined by whether AI strengthens the relationships, judgement and reflection on which great teaching has always depended.

Dr Melissa Bond is a research fellow at the National Institute of Teaching