Education and AI

Project background

Data and digital technologies have long played a role in education in the UK. For example, information on students’ learning and attainment is used not only to assess progress, but also to inform school performance rankings and policymaking. Schools’ use of digital technologies, primarily for teaching and learning, spiked due to remote learning during the COVID-19 pandemic.

The use of artificial intelligence (AI)-driven technologies in education is varied.  ‘Narrow’ AI –task-specific systems trained with a dataset that is curated for a particular purpose – is used for administration tasks such as analytics, safeguarding, classroom management, SEND education and some personalised learning.

Teachers’ use of general-purpose AI or generative AI (such as commercial and non-education specific products like ChatGPT, Microsoft Co-Pilot or Google Gemini) for lesson planning and to support marking and assessment is mostly informal, despite the hype and investment.

The use of these technologies raises questions about their impact on outcomes, not least in relation to the accuracy of their outputs. Meanwhile, the development of AI  technologies for use in education is in its infancy, with a handful of products emerging. These are yet to be evaluated for accuracy, efficacy and efficiency in the classroom.

While AI and education-specific AI technologies have the potential to support teachers with their planning, teaching, administration and time, the benefit for pupils and students is yet to be proven clearly.

Furthermore, the risks and potential harms from their use are yet to be fully understood, and many are unique to these new technologies.  Previous harms caused by algorithms, narrow AI and the use of data-driven technologies within education – such as the failure of the A-level grading algorithm in 2020, or the data protection and data exploitation risks associated with commercial learning platforms such as Google Classroom and ClassDojo – highlight the need for safeguards to be established to ensure that the benefits can be realised without the negative impacts being experienced by students and staff.

As data-driven and AI technologies develop, and their procurement in education increases, more research is needed to assess their impact on teaching practices and administration, and on students’ development and attainment.

Project overview

The Ada Lovelace Institute and the Nuffield Foundation are collaborating on research into the role of AI and data-driven technologies in education.

In 2023 we submitted a response to the Department for Education’s call for evidence on generative AI in education.

A landscape review of AI in education in UK state-funded primary and secondary schools will be published in early 2025. This will launch the next phase of this work, with the aim of evidencing the opportunities, benefits and risks of AI in education. This could be in relation to teaching and learning, marking and assessment, or careers guidance.

Furthermore we will explore the evaluation of AI technologies being used in schools. How can we ensure they are pedagogically and technically fit for purpose, and that the outcomes they have are beneficial and can be proven?

Image credit: JohnnyGreig