The government’s go-to education research body has opened a £2.5 million research fund to fill an “urgent evidence gap” on how tools like ChatGPT have impacted the way pupils learn. The Education Endowment Foundation (EEF) said the project will aim to understand how generative AI (GenAI) influences learning processes and outcomes. The foundation is especially interested in how the technology may lead pupils to “offload” thinking tasks such as recall, planning, reasoning or drafting, onto the tools. The project will look at whether it supports deeper learning, or whether it weakens pupils’ engagement, reduces their ability to remember what they have learned, or risks making them overly reliant on AI. Another key aim is to explore how AI affects different learners, particularly those from disadvantaged backgrounds. The EEF will use the fund to commission several studies which will examine a range of educational uses of AI, such as summarising content, solving problems, essay planning and receiving feedback. It will also investigate how AI use impacts outcomes, such as knowledge acquisition, working-memory and problem-solving, as well as motivation and resilience. ‘Robust evidence’ needed More than two-thirds of 13 to 18-year-olds are using AI to support their literacy and learning, according to a recent National Literacy Trust report. Becky Francis, EEF chief executive, hoped the programme would create a “more robust understanding” of AI’s impact on learners. “There has been a lot of speculation about the potential of GenAI tools like ChatGPT to change teaching and learning, for better or worse. But the collection of robust evidence on their actual impact on learning has barely begun – especially for learners under age 16. “As these tools become more widely embedded, there is an urgent need to develop a much clearer understanding of whether and how GenAI supports or hinders the outcomes we want for learners. “And we need to avoid any unintended consequences in widening inequalities for different groups of learners.” Research teams are being invited to take part. They will have until 30 June to submit their expressions of interest. The research is expected to begin this year, with first findings expected in 2027.