Frequently Asked Questions – Estyn and Artificial Intelligence
At Estyn, we are beginning to use artificial intelligence (AI) carefully and responsibly to support aspects of our work. AI is a tool to help inspectors, not to replace them. All inspection judgements are made by inspectors, based on evidence and professional expertise. AI helps us with tasks like analysing information, producing learner-friendly versions of reports, saving time and making inspection findings more accessible. Inspectors review every AI-generated output to ensure accuracy, fairness and clarity. We have outlined the principles underlying our approach toward the use of AI in our Ambitions for AI statement.
We are currently using AI in specific, practical ways to support and enhance our work, while keeping inspectors’ professional judgement at the heart of everything we do. For example, we have been piloting learner-friendly versions of inspection reports in the primary sector. These use AI to produce clear, age-appropriate summaries so that pupils can better understand the strengths of their school and areas.
At the same time, we are exploring the potential of AI across other areas of our work. This includes looking at how it might help us analyse large amounts of information more efficiently or present findings in different accessible formats for learners and providers.
Importantly, we only use AI where it adds real value to our work and where it is of benefit to learners, staff, and providers of education and training across Wales. Every output from AI is reviewed and approved by inspectors, ensuring that the quality, accuracy, and validity of our inspections and evaluations remain as high as ever.
No. AI will never replace the professional judgement and expertise of inspectors. It is only used as a support tool, for instance to summarise information or present findings in more accessible ways. All inspection outcomes are based on the expertise and professional evaluation of our inspectors.
Inspectors always review and refine AI outputs. We are very aware that AI can sometimes produce inaccuracies or reflect biases in the data it has been trained on. This is why we use AI to assist with tasks, not to make inspection decisions. We ensure that human oversight is central to our use of AI. Inspectors always check AI outputs carefully. If AI generates something inaccurate or misleading, it is corrected before being used. The responsibility for accuracy and clarity rests with the inspector, not the AI.
No. Inspection judgements remain entirely in the hands of inspectors. AI has no role in deciding on judgements or evaluations; inspectors may use it to support the process, for example by drawing together themes from evidence gathered through first-hand inspection activity.
We use AI to enable us to communicate our findings more widely. For example, producing simplified versions of inspection reports for learners would be very difficult to do manually on a large scale. We also use AI to improve the efficiency of our work. For example, using AI to draw together initial themes from inspection evidence provides more time for inspectors to engage in the most important aspects of the inspection process, such as talking to learners, staff and leaders.
We take data protection and confidentiality very seriously. AI is only used with information that we can process securely. We never use AI tools in a way that would put confidential school data at risk. More information about our handling of personal data can be found in our privacy notice.
Inspectors always review and edit AI outputs before sharing them. For example, reporting inspectors edit AI generated learner-friendly reports for accuracy, tone, and suitability. AI provides a draft, but inspectors ensure the final version is clear, correct, and appropriate. All our reports are subject to rigorous quality assurance processes: Estyn’s arrangements for assuring the quality of inspections
No. The validity of inspection reports comes from the evidence gathered and the professional judgement and expertise of inspectors. AI is a tool that can help us communicate findings more effectively, not a replacement for our inspection methods.
There are a number of benefits for providers and learners resulting from our use of AI. For example, AI is helping us create reports that are more accessible for learners, making inspection findings more meaningful to them. This helps learners engage with what’s going well in their setting and how they can play a part in improvement. In addition, using AI to support their work enables inspectors to spend more time gathering the views of learners, staff and leaders resulting in a more supportive inspection process and accurate reports.
We are exploring how AI might be used responsibly in other parts of our work, but always with care, transparency, and human oversight. Any expansion will be gradual and based on evidence of what works, what is safe, and what supports schools and learners best.
We don’t inspect the use of AI as a separate aspect. Instead, where providers are using AI, we consider how well it supports effective teaching and learning. Our focus is always on the quality of education and the impact on learners.
As the use of AI in wider society grows, we will increasingly look at how providers are helping learners to understand its opportunities, risks, and challenges. This includes how well learners are being prepared to use AI in an ethical, responsible, and safe way.