EEF opens research call on how generative AI affects student cognition

Research teams have until 30 June to submit expressions of interest for studies on cognitive offloading, learning, and GenAI use among 13- to 15-year-olds in England.

The Education Endowment Foundation is commissioning research into how generative AI use affects cognition and learning among 13- to 15-year-olds in England.

The Education Endowment Foundation has opened a research commissioning call to examine how school-aged learners’ use of generative artificial intelligence affects cognition, with expressions of interest due by 30 June 2026.

The call focuses on cognitive offloading, where students delegate cognitive processes such as recall, planning, reasoning, drafting, evaluation, or problem-solving to a generative AI system. The Education Endowment Foundation is seeking research teams to test when that use supports learning and when it may weaken or distort the learning process.

The invitation to tender opened on 8 June 2026 and is focused on learners aged 13 to 15 in England. Research teams must submit questions by 9am on 18 June, with responses due to be published on the ITT webpage on 23 June.

The Education Endowment Foundation expects to commission several research projects through the call. Shortlisted applicants will be invited to submit full proposals on 16 July, with full proposals due on 2 September and selected applicants expected to be notified on 16 September.

The research is being commissioned as generative AI tools become part of students’ everyday learning. The Education Endowment Foundation cites a 2025 National Literacy Trust report in which two in three 13- to 18-year-olds reported using generative AI to support literacy and learning.

Focus on cognitive offloading

The Education Endowment Foundation wants proposals that test how generative AI affects learning and cognition, rather than research that only describes how students are using tools such as ChatGPT, Claude, or Gemini.

The call is particularly focused on general-use generative AI tools, including large language models, rather than education-specific products such as AI tutors or lesson planning assistants. Researchers will need to show how safety issues, including data privacy, protection, security, intellectual property, and safeguarding, will be addressed in the study design.

The Education Endowment Foundation says generative AI can be used by students to summarize information, expand or rephrase concepts, provide worked examples, support guided reasoning, offer formative feedback, and help with planning, monitoring, and evaluating learning.

The research will look at cognitive outcomes such as short-term memory, working memory, attention, critical thinking, metacognition, idea generation, knowledge acquisition, and problem-solving. Secondary outcomes could include motivation, self-efficacy, and resilience.

Applications from research teams

The call is open to legally constituted organizations, not individuals. Research teams do not need to be part of the Education Endowment Foundation’s existing panel of evaluators.

Applicants may apply as a single research team or as a consortium with complementary expertise. The Education Endowment Foundation is looking for teams with experience in experimental education studies, a grounding in learning theory, and, ideally, experience in AI in education.

Organizations do not have to be based in England, but teams based elsewhere will need to show they have the partnerships and knowledge required to work with English schools.

The first-stage expression of interest should be no more than 2,000 words, excluding references and CVs. Full proposals invited at the second stage should be no more than 7,000 words, excluding references.

Experimental evidence and school relevance

The Education Endowment Foundation says competitive proposals should use research designs that can generate robust evidence, including strong causal inference, a credible counterfactual, transparent research design, and replicability. Purely descriptive or correlational studies are unlikely to be competitive.

Projects may be lab-based or classroom-based, although the Education Endowment Foundation says it will prioritize studies with the greatest potential to inform pedagogy. Randomization may take place at setting, classroom, or learner level, depending on the research question.

The Education Endowment Foundation is also open to proposals with staged designs, including studies that first examine a short-term cognitive outcome before testing longer-term outcomes such as knowledge retention or transfer to other domains.

The call asks research teams to consider effects on socioeconomically disadvantaged students, including differences in access to generative AI and the support students receive when using it.

Commissioned projects are expected to produce research reports for publication on the Education Endowment Foundation website, with participating schools anonymized. The Education Endowment Foundation also expects peer-reviewed study plans to be published.

Set-up meetings are scheduled for 20 September to 2 October, with projects due to go to the Education Endowment Foundation’s Grants Committee in November. Project commencement is expected in November 2026, pending Grants Committee sign-off.