AI in Assessment: 6 principles for schools to follow 

Provided by @Laura Bailey, Head of Moderation at BESA member, Pobble

The use of AI in schools is changing the conversation around assessment, bringing new opportunities alongside some important questions about how the technology should be used. 

What should remain firmly in the hands of teachers? How can schools make sure AI assessment tools genuinely improve teaching and learning? And what should they look for when choosing who to work with? 

These were among the topics explored during a recent AI in Assessment panel hosted by BESA member, Pobble, and chaired by Al Kingsley MBE, CEO of fellow BESA member NetSupport.  

The panel of experts covered everything from teacher workload and professional judgement to moderation and the practical use of AI in assessment. 

Following the discussion, we asked Laura Bailey, Head of Moderation at Pobble, to share six principles she thinks schools and trusts should keep in mind when exploring AI-supported assessment. 

1. Start with the problem you want AI to solve 

Is it workload? Inconsistent assessment? A need for better insight into progress? 

If you’re clear about the problem you’re trying to solve first, you can decide whether AI is the right solution and what you would expect it to improve. 

Otherwise, there’s a risk of introducing yet another tool for teachers to manage without being clear about the value it adds. 

2. Decide what should remain human 

One particularly useful idea raised during the discussion was ‘non-negotiable human thinking’ – the things we don’t want to outsource to AI. 

AI might identify patterns in spelling, punctuation or particular writing skills, but only a teacher can fully understand the context, e.g., how independently something was written, what support was provided and what the pupil was trying to achieve. 

There’s also the human response. Was the writing funny, moving or surprising? Did it connect with its audience? 

The key is knowing where AI can add useful insight, and where context and professional judgement still need to lead. 

3. Make sure insight leads to action 

More assessment data doesn’t automatically improve teaching.  

AI can help identify a skill a pupil has consistently struggled with, progress that might otherwise be overlooked or a pattern across a class – but the important question is: what happens next? Does the teacher need to revisit a skill? Adapt the next lesson? Speak to a particular pupil? 

The value comes from what teachers can do with the information gained. 

4. Bring teachers with you 

Introducing AI can represent a big change in practice. Teachers need to understand what a tool can and can’t do, and have the confidence to use it effectively and question its outputs. 

That requires continued professional development, clear expectations and continued opportunities for professional discussion and moderation. 

It also shouldn’t all sit with one nominated ‘AI champion’. Schools need a shared approach where everyone understands how the technology should (and shouldn’t!) be used. 

5. Understand AI safeguarding and data protection 

Using AI in schools introduces additional considerations around data protection, safeguarding, bias and pupil privacy, particularly when tools are being used to analyse pupils’ work or inform assessment. 

How has the tool been tested? How consistent are its judgements? How has potential bias been considered? How will pupil data be handled? 

Alongside established processes around data protection, DPIAs, safeguarding and procurement, schools should make sure they understand the specific risks and safeguards associated with the AI tool they are considering. 

6. Look at the organisation as well as the product 

Trust is an important part of choosing any supplier – particularly when choosing an AI or EdTech supplier whose technology handles pupil data or informs assessment. 

Can the supplier explain how their technology works and provide evidence to support it? How will they protect your data? What support will staff receive? And do they understand the responsibilities that come with working in education? 

As technology develops, that sense of trust will only become more important.  

— 

Pobble is Backed by BESA, which means committing to BESA’s Code of Practice and its standards around quality, transparency, safeguarding and responsible conduct.  

When schools are considering new suppliers – particularly in fast-developing areas such as AI – BESA membership offers a useful marker to look out for. After all, if it’s built for schools, it should be Backed by BESA.