EEF opens £2.5m fund to study AI’s effect on learning

TL;DR:

• The Education Endowment Foundation has opened a £2.5m fund to address an “urgent evidence gap” on how generative AI affects how pupils learn.
• It will examine whether tools like ChatGPT support deeper learning or let pupils “offload” recall, planning and reasoning at a cost to memory and engagement.
• With over two-thirds of 13- to 18-year-olds already using AI for learning, first findings are not expected until 2027.

Britain’s schools are absorbing generative AI faster than anyone can measure its effects. The Education Endowment Foundation, the government’s main education research body, is trying to close that gap with a £2.5m programme to find out whether tools such as ChatGPT genuinely help children learn — or quietly undermine it.

Studying the offloading question

The fund’s central worry is cognitive offloading: when pupils hand tasks like recall, planning, reasoning or drafting to a chatbot, does understanding deepen or atrophy? The EEF will commission several studies covering common classroom uses — summarising content, solving problems, planning essays and receiving feedback — and measure effects on knowledge retention, working memory, problem-solving, motivation and resilience. A particular focus is how AI affects disadvantaged learners, where the foundation’s remit and the risk of widening inequality are sharpest.

The urgency is well founded. A recent National Literacy Trust report found more than two-thirds of 13- to 18-year-olds already use AI to support literacy and learning, well ahead of any robust evidence on whether they should. “The collection of robust evidence on their actual impact on learning has barely begun — especially for learners under age 16,” said EEF chief executive Becky Francis.

The move complements the government’s wider skills agenda, including AI bootcamps aimed at protecting young workers — evidence on schooling being the upstream half of the same question.

Looking forward

Research teams have until 30 June to register interest, with studies starting this year and first results in 2027. That timeline is the catch: classroom AI use is racing ahead while the evidence to govern it remains two years out. Schools, in the meantime, are left improvising policy without the data the EEF is only now setting out to gather.