Artificial Intelligence - Estyn

Artificial Intelligence


Our Ambition

Setting a safe environment for artificial intelligence (AI), we will use AI to support more efficient working. The efficiency gains and insights will free up resource to further enhance our services. Having robust governance arrangements in place will help offer clear guidance to our staff and stakeholders.

We will realise our ambition though engaging with our stakeholders and partners and create the space and environment for discussions to flourish on a range of important areas. Our position is stated below which will continue to evolve as we continue to engage with AI expertise and discussions with our partners.

AI Guiding Principles

Purposeful Use:
We will not adopt AI for its own sake. Its use will be purposeful—supporting and enhancing our work to drive improvements for learners, for Wales.

Responsible Governance:
We will update our Data Protection Impact Assessments (DPIAs) regularly to ensure full compliance with data protection legislation and best practice.

Strong Oversight:
AI will contribute to our work but not replace our work. All AI-generated outputs will be reviewed and approved by staff through our established quality assurance processes.

Adaptive and Vigilant:
As AI evolves, so will our approach. We will explore emerging opportunities while remaining alert to potential risks—especially in education, training, inspection, and wider organisational activity.

AI and data

Compliance with legislation and best practice is paramount. Safe and appropriate use of our data is our principal consideration. We will achieve this by prioritising safe data use over innovation.

As part of our data protection impact assessments, we will thoroughly research new AI tools, licenses and privacy policies before we begin to trial and experiment. Inspection related information will be kept safe and secure.

AI and inspection

As we inspect providers who themselves are exploring AI, our position on AI mirrors our broader approach to inspection – we do not prescribe any preferred methods or tools (including AI). Our inspections remain focussed on the impact these tools may, or may not, have on leadership, well-being, teaching and learning.

AI, bias and transparency

Our inspection and other work are subject to rigorous quality assurance processes. As we continue to develop the use of AI, we will be transparent about how we have used AI in any of our processes.

Our quality assurance also guards against bias as part of statutory duty to ensure that inspectors carry out inspections of consistent high quality.

AI, equality and the Welsh language

Our adoption and use of any AI tool will be made pursuant to the Welsh Language Standards (No. 2) Regulations 2016. The Welsh language will be entrenched in our planning from inception stage. We will test any new AI developments using our integrated equality impact assessments.

We will ensure protected characteristics under the Equality Act 2010 are considered.

AI and the impact of poverty

We recognise the impact that socio-economic inequity has on educational outcomes.   Access to AI tools and the digital skills needed to use them effectively may be limited for individuals and communities experiencing poverty.

 We will engage with our partners to ensure that we contribute to the promotion of inclusive access, digital literacy, and fair opportunities for all learners and practitioners, regardless of socio-economic background.

AI and the environment

As AI adoption grows, the demand for energy-intensive infrastructure continues to rise, exacerbating climate change and resource depletion. We will use AI prudently and explore ways to minimise harmful effects as part of effective environmental management e.g. utilise data centres using renewable energy sources and use energy efficient AI tools where possible e.g. GPT ‘mini’.

Our priorities for 2025 – 2026

During the coming year we will focus our work in three main areas:

  • Estyn’s AI work: We will continue to safely experiment with how AI can support efficiency and enhance our own work, including incorporating areas we have been piloting into our business as usual
  • Education and training AI rapid review: We will enhance our collective understanding of how education and training providers are using AI and share effective practice
  • International collaboration: We will work with international partners to evaluate how AI can enhance inspection work, drawing on international best practice

AI and our governance

We will ensure our governance arrangements have appropriate oversight of our delivery against this AI ambition. In addition, we have established an AI oversight group, with cross organisation representation to directly lead on our work in this area. As we realise the potential of AI in our work, we will commit to review our approach against our 4 key areas.