The AI revolution: why equity must drive the future of tutoring

At Action Tutoring, our vision is simple: ensuring that a young person’s background does not dictate their future. For over 14 years, we have seen the transformative power of a dedicated tutor through engagement and a targeted curriculum.

While our volunteer model has a proven, life-changing impact, the sheer scale of the attainment gap requires us to think bigger. There are currently 2.2 million pupils in receipt of Free School Meals in the UK today who could benefit from our support.

That is why, in autumn 2026, we are launching an ambitious new pilot of AI tutoring, to bridge this gap, independent of the work announced by the Department for Education in January 2026.

We’re interested in how an AI tutoring model which keeps ‘humans-in-the-room’ can combine the ‘personalisation power’ of Generative AI with the ‘engagement power’ of human relationships.

Personalisation at scale

Research consistently shows that ‘personalisation’ is the secret weapon of tutoring. For that reason, the ability to find a pupil’s specific learning gap and bridge it with tailored feedback is something it’s impossible to replicate in a standard classroom.

Generative AI offers extreme personalisation. In essence, when harnessed correctly, it can scaffold instruction and provide instant feedback at a level of precision that was previously a luxury of those who could afford private tutoring, or those who could access it through an intervention like ours. We believe AI tutoring has the potential to transform attainment for those facing disadvantage, which in turn could substantially reduce the attainment gap.

Closing the new digital divide

Innovation in education is all too often a luxury of the rich. Currently, private school pupils are benefiting from Generative AI at triple the rate of their state-school peers. Action Tutoring refuses to let AI become yet another barrier that widens the gap between the ‘haves’ and the ‘have-nots.’

Therefore, by piloting an AI-delivered model, we are democratising access to the world’s most advanced learning tools. Our commitment to equity underpins the moral argument for Action Tutoring to pursue this pilot.

The ‘human-in-the-room’ model

A decade of evidence-building has taught us that for young people facing disadvantage, tutoring is as much about engagement and relationships as it is about pedagogy.

We aren’t going to sit pupils in front of a screen and walk away. Our pilot is built on a ‘human-in-the-room’ design. And it’s not just us that recognises the importance of this: three of the four published randomised controlled trials about AI tutoring specifically include a human element in their design and all show evidence of promise.

In practice, for us, this means:

• The AI tutor delivers the Action Tutoring-tailored curriculum we’ve refined over many years and provides real-time feedback.
• The Tutoring Coordinator (a synchronous human presence) supports pupil engagement, manages the technology, and provides the relational encouragement that no algorithm can replicate.

We are testing the optimal human support required to make AI effective for disadvantaged pupils, rather than looking at the technology in isolation.

A note on safeguarding

Our ability to uphold our high standards of safeguarding has been included as a non-negotiable in the criteria that we are using to assess AI tools. We have built up 14 years of safeguarding expertise which will also be integrated into the system and our approach. This includes retaining our human presence as a connection to both the pupils and the teachers on the programme. We will be carefully managing risks and including additional safety considerations in the design phase. This may look different for different tools but could include ‘screening’ or post-session review options. We’ll also be building in guardrails into the prompts so that any tool we use has clear boundaries about what it can discuss and how to manage any difficult issues that pupils raise. Our wealth of training and resources we currently use for our tutors can also be used to train our AI tool.

And what about our existing programme?

We know that for many of our supporters, the ‘human’ in tutoring is what matters most and we agree.

We are not abandoning our volunteer-delivery model. Our volunteers are the heartbeat of Action Tutoring. This AI pilot is an addition to our portfolio, not a replacement. Specifically, we believe AI could significantly expand our reach, allowing us to support thousands more young people in budget-constrained schools who we currently cannot support. Our commitment to our volunteer community remains as steadfast as it was 14 years ago.

What’s next?

This autumn, we begin our small-scale, mixed-method pilot in around eight secondary state schools across the country. Delivery will follow the same format of our existing high-impact, evidenced-based model and use a safe, Generative AI tool to deliver the tutoring.

We will be collecting both qualitative and quantitative data to understand the optimal structure of the human support required to deliver our curriculum-tailored programme to disadvantaged pupils, through an AI tool.

Additionally, we will begin to build an understanding of the efficacy of AI tutoring for our target population of pupils from both an academic progress and a pupil experience perspective. We aim to test and learn what AI tool functions are required to deliver our model of tutoring.

We are an evidence-based organisation. We won’t rush; we will move thoughtfully and build things that work. In the long run, our goal for 2027 and beyond is to discover how we can unlock the potential of thousands more disadvantaged young people every year.

The AI revolution is happening. At Action Tutoring, we’re making sure it’s an equitable one.

If you’re a school leader who would like to be involved in our innovative new pilot, register your details below.