Artificial intelligence (AI) can mark our children’s homework, assess what they need to learn (and the best time to learn it), or even detect when they are cheating in exams. But what does an education system embedded with artificial intelligence look like?

Nesta is involved in lots of work seeking to maximise the public benefit of AI. However, across government, industry and academia, relatively little attention has been paid to how AI could affect our education system (compared with, say, healthcare or e-commerce).

We’ve begun a research project exploring the potential of AI to impact on our education system. But how do you research something that hasn’t really happened yet?

Given that so much about the future of AI and our education is uncertain, we will be adopting ‘Futures Tools’ to explore different scenarios for the future. We’re not trying to predict the future. But instead, we hope to understand what needs to happen now to make a positive future for our education system more likely.

Before considering potential future scenarios, we have spent the last two months examining the present. This blog gives an overview of what we’ve found out, points towards a future direction for the project, and invites you to share your ideas for how our exploration should continue.

Over the last two months, through desk research and interviews with experts, we have set out to identify and review:

1. Existing academic and grey literature on the subject
2. Existing AI being deployed or developed for use in our education system now.

We’ve discovered a relatively modest - but fast-growing - bank of academic literature focusing on artificial intelligence in education. As the literature on AI in education grows, we can also see its scope widening.

Early academic literature was typically focused on how AI could be used to solve ‘Blooms 2-Sigma Problem’ and replicate the ‘gold standard’ of education: 1-2-1 tutoring. However, academics, researchers and technologists are now describing experiments with AI focused on whole range of different elements - from enabling collaboration between peers to assessing complicated skills like creativity.

We found a similar trend when looking at AI products. Although most of us think of ‘adaptive learning platforms’ (software programmes that use AI to personalise the content and pathways that learners follow based on their habits, progress and interests) when we think about AI in education, there is a much wider range of AI applications in our education system - from managing school inspections to automating feedback for learners.

Below are the AI applications we have discovered grouped AI into three categories - learner-facing, educator-facing and system-facing.

Although relatively young, these applications have great potential to help address some of our education system’s most persistent and stubborn problems. Below we have outlined nine challenges that AI is well-placed to tackle:

1. Teaching overwhelmed by administration
2. Inflexible progress through the education system
3. Difficulty of improving teacher practice
4. Failure to defeat entrenched social immobility
5. Difficulty of teacher recruitment and retention
6. Unsuitability of current curriculum for future society and economy
7. Difficulty of facilitating alternative pedagogy
8. Culture of assessment inhibiting teaching and learning
9. Homogenisation and lack of personalised learning

Of course clever technology doesn’t necessarily mean clever solutions. In the next stage of the project, we will unpick some of the issues that will make the impact of AI in our education system more positive or negative, or both.

We’re going to explore the future of AI in education through a series of future scenarios from the perspective of educators and learners. Rather than serve as predictions, these will help us consider alternative futures and what choices might be made today to help generate positive outcomes. We will look out to around 2035 as this gives sufficient time to imagine substantive change without too much speculation.

The future of AI and education is a substantial topic so we will focus on those aspects that will have most impact on educators and learners, and the future of which is most uncertain. Themes for exploration include who controls the data that underpins AI and how these technologies might fundamentally change the assessment of students, perhaps with moves away from milestone exams and toward continuous assessment of many more capabilities.

If there’s an issue you think we should consider, an area where we should concentrate our efforts or if you want to find out more - please get in touch via [email protected].