Exploring Multi Academy Trust approaches to Artificial Intelligence

This project explored how multi-academy trusts (MATs) have approached the challenges and harnessed the opportunities of AI. It also looked to understand how AI tools have been implemented, evaluated and cascaded across schools to support better teaching and learning.

Why this project is important

General-purpose AI tools have the potential to transform the ways school operate through more intelligent data collection, more accurate assessment setting and marking, personalised and adaptive learning, and improving teachers’ administrative workload. By 2024, 57% of teachers were using tools like ChatGPT, more than one in ten had used an AI tool in a lesson, and 31% had not used AI at all.

Despite the rapid adoption of AI, significant barriers exist to exploiting the benefits more broadly, including a lack of evidence of impact, national guidance, and oversight. The Department for Education has taken steps to overcome these barriers. However, the EdTech and AI market and capabilities are evolving very quickly which makes it difficult to ensure that guidance and safeguards for the sector are sufficiently up to date.

This project sought to examine current practices and identify where further support or intervention is needed in order to improve the system’s response to AI.

What the project involved

The research team conducted some rapid work – scoping the current policy and research context, and convening two roundtables with representatives from MATs, policy specialists, and educational organisations.

The first roundtable focused on the ways in which AI is currently being used and asked:

• How are MATs currently using AI?
• How is AI implanted across trusts?

The second roundtable focused on governance and ethical considerations of using AI in academy trusts and asked:

• How are MATs assessing the effectiveness of the use of AI?
• What are the decision-making processes that MATs adopt when deciding on an AI strategy?
• How are MATS managing legal and ethical considerations?

What the project found

• Adoption of AI is uneven. MATs use AI to support teaching, learning and administration, and its presence in everyday digital tools makes some level of use inevitable. However, how systematically and strategically AI is adopted varies widely between trusts and between schools within the same trust.
• AI has the potential to reduce workload and support learning. AI can reduce lesson-planning time, support communications, and enable more personalised learning, particularly for pupils with SEND or EAL. However, MAT leaders cautioned that efficiency gains do not always translate into reduced workload, and concerns remain about bias, reduced human interaction and impact on wellbeing.
• Some MATs take bottom-up approaches to AI adoption that rely on teacher-led experimentation and feedback, often through small pilots, ensuring policies reflect classroom realities and preserve autonomy. This approach fosters innovation but requires complementary top-down oversight for consistency and safety. Other MATs take a predominately top-down approach and enforce strict approval processes.
• MAT leaders emphasised that AI adoption must align with clear educational goals.
• MAT leaders described having to navigate a “wild west” of AI products – and felt that decision making around AI was hindered by insufficient national direction, guidance and evidence.
• Evaluating impact and managing risks were major challenges. While larger MATs may conduct evaluations or negotiate with providers, smaller trusts often lack the capacity to make fully evidence informed choices. MATs mainly rely on surveys, feedback and usage data to assess effectiveness, which makes it difficult to isolate AI’s impact on outcomes.
• Legal, ethical and equity issues such as data protection, provider transparency, and the digital divide are central concerns, with risks that AI adoption could exacerbate existing inequalities without stronger coordination and support.

Project Recommendations

Recommendation 1: The Department for Education should continue to evaluate the effectiveness of Edtech products and their use through the Edtech Evidence Board project. In addition, the sector should be incentivised to share their own approaches to evaluating the products they are using, with larger trusts well placed to support other schools in the system.

Recommendation 2: Create research informed guidance on developing AI and digital literacy for education providers AND for initial teacher training (ITT) programmes.

Recommendation 3: Larger trusts should lead networks of support working with both smaller trusts and individual schools with the Department for Education considering ways that this could be incentivised. The Department for Education and the Department for Science, Innovation and Technology should work across government to ensure that disadvantaged communities (either through economic circumstances or their location) are not left behind as technology progresses, ensuring access to devices and high-speed reliable internet.

Recommendation 4: The Department for Education should consider the merits of providing the key information that trusts and schools will need for completing DPIAs for the more widely used AI products while being mindful of the fact that the process of producing a DPIA provides a structured approach for data controllers to consider their individual circumstances.