Experimentation to evidence: Assessing generative AI impact in schools

This project will investigate how schools in England can systematically evaluate the impacts of generative AI tools on learning, wellbeing, and equity.   

Why is this important?

Schools across England are rapidly adopting generative AI, with 44% of teachers and 80% of 13-18 year old students using AI tools in education. However, adoption is largely driven by enthusiasm rather than evidence, raising concerns about equity, student effort, privacy, and learning outcomes. A lack of systematic frameworks to evaluate AI’s impact leave schools unable to distinguish effective practices from harmful ones. 

What does it involve?

The research team will work with schools involved in the Shape the Future Coalition, a network of over 1000 schools actively exploring AI uses across the UK. The project will address four research questions: 

1. What impacts are emerging from Generative AI use across multiple dimensions in Coalition schools? How does this vary by school context, implementation approach, and student population?  

2. What enables schools to generate credible evidence about AI impact? Which existing evaluation approaches produce actionable evidence, what are the barriers preventing systematic assessment, and what are the learnings from schools’ experience measuring GenAI impact? 

3. How should an Impact Evidence Framework be designed for diverse school contexts? 

4. How do school leaders perceive students’ role in GenAI impact evaluation?  

The research will be completed in five stages:   

1. Conducting a rapid evidence review and in-depth comparative analysis of seven schools undertaking structured AI implementation projects. 

2. Mapping the broader landscape of AI use and evaluation practice through a survey distributed to 70+ AI leads. 

3. Developing the Impact Evidence Framework. A sequence of structured workshops, each involving 15-20 educators, will inform development, to define impact dimensions, review evaluation approaches, and assess framework usability. 

4. Piloting the full framework in three or four schools. 

5. Synthesising the findings and producing practical tools, such as decision trees, outcome indicators, data collection templates, analysis guides, and worked case studies illustrating the Framework application in different contexts. 

How will it make a difference?

Key outputs such as public reports, toolkits, and policy briefings will be disseminated through professional associations and conferences. The team aim to lay the foundations for further work that would look to scale the implementation and tracking the use and impact of the Framework over time.