The visible reminder of invisible learning

 Author

Our sector is working hard to make sense of, and to address, the impacts of AI on teaching, learning and assessment, at a point when a recent survey saw 38 per cent of students admit to submitting work which they say they don't understand.

In this context, it seems essential that our approaches to the design of learning prioritise assessment strategies that remain valid in an AI-enabled environment, to retain the visibility of that learning and to encourage students to develop and demonstrate those key skills and capabilities which AI alone cannot replicate.

And so, recognising that AI forces us to have a much clearer idea about what assessment is trying to measure, we at Birmingham City University are introducing a new GenAI-integrated assessment framework for academic year 2026-27. We originally developed this framework to support the work of our Department of Education, but, following positive feedback from the BCU AI Working Group it was agreed it should be rolled out across our institution.

This framework is structured around five pillars: purpose and learning outcomes (clarifying what learning must be independently demonstrated and where AI may support – or should be limited to protect – core academic capabilities); authenticity, professional relevance and real-world practice; ways to diversify how students demonstrate learning through assessment modes and evidence types; the transparency and accountability of AI use and the clarity of expectations around that; and the embedding of academic integrity, equity and accessibility.

In working to support our colleagues in the work of redesigning assessments and curricula in response to a world fundamentally impacted by the rise of generative artificial intelligence, we've also come to recognise the massive variation in those colleagues' levels experience and capability in the use of AI. Some have never engaged with AI at all, some are highly sceptical about it, others are extraordinarily enthusiastic and creative in its deployment. It has become very clear to us that we need to bring our whole institution up to a minimum standard to incorporate AI across the board, and that this is something we need to do because that's the way the world is going and where we need to be.

If we're in the business of preparing students to enter the workplace and to succeed there, then we need to change – just as our society is changing – and that can sometimes be challenging and even uncomfortable. As we've found conducting focus groups across our university, different individuals and different disciplines bring with them a broad and rich variety of experiences and expectations – including, crucially, their expectations of academic and professional integrity.

And so, it's been vital to stress that this isn't simply about showing students how to write an AI prompt (the technology is after all designed by its nature to be user-friendly). It's about supporting the development of their own critical literacy and agency. This is why we need our students to learn to be transparent in their uses of AI – not just in declaring that they've used it, but in demonstrating how and why they've done so, and the ways in which they've thought about it, and interpreted and interrogated it. It's about students developing strategies through which to use AI, rather than through which to be used as passive vehicles for the technology.

To accept that artificial intelligence has the potential to do great good does not mean that we should ignore the big issues around its environmental, economic and societal impacts. It doesn't mean that we should stop questioning the ways it may contribute to the concentration of power, wealth and influence in our world.

That's why we need strong guidelines and policies to safeguard the uses of AI, both in higher education and far beyond. The work of universities in this area can help to lead on this.

After all, if every student were to use the same AI tools to write their essays, we'd end up caught in the monotony of a single, standardised academic voice. That's certainly not what higher education should be about. This is why it's so important to shift the focus of assessment and learning away from that traditional emphasis on the final assessment point's finished product towards the use of process logs and portfolios which articulate evidence of thinking and understanding, of our students' engagement with ideas and their critical development. The process of assessment should be a process of learning and should demonstrate and promote that process. If every student relies on the same AI tools in the same ways, we risk creating cohorts of academic gingerbread men, neatly formed, technically competent, and remarkably similar. Higher education should be concerned with cultivating distinct voices, perspectives, and ways of thinking, not reproducing uniformity at scale.

Assessment should always capture and stimulate our students' capacities for understanding, reasoning, judgment and decision-making. In this age of artificial intelligence, we can continue to achieve this by ensuring that, through the processes of assessment, we make learning visible, traceable and defensible. And in doing so, we can remind ourselves, our stakeholders and our critics, the purpose and value of higher education – as our students chart and witness the extraordinary and unique transformations it can inspire.