This blog was kindly authored by Dr Emma Ransome, Academic Lead for Teaching and Learning at Birmingham City University.
Assessment is fast becoming a central focus in the higher education debate as we move into an era of generative AI, but too often institutions are responding through compliance and risk-management actions rather than fundamental pedagogical reform. Tightened regulations, expanded scrutiny and mechanistic controls may reassure quality assurance systems, but they run the risk of diluting genuine transformation and placing unsustainable pressure on staff and students alike.
Assessment is not simply a procedural hurdle; it is a pivotal experience that shapes what students learn, how they engage with content and what universities and employers prioritise as valuable knowledge and skills. If reform is driven through compliance, we will miss opportunities to align assessments with the learning needs of a graduate entering the gen-AI era.
Generative AI as a lens on existing assessment practices
Tools such as ChatGPT pose real questions for assessment validity and integrity, but they also expose long-standing systemic tensions. The UK Quality Assurance Agency’s sector guidance on generative AI highlights that assessment practices developed under traditional paradigms are under strain in a world where students can access powerful text-generating tools online.
The QAA’s advice emphasises that institutions need sustainable assessment strategies that move beyond detection alone and calls for principled redesign rather than reactive policies. Increasingly, generative AI is being referred to as a ‘wicked problem’, one not amenable to simple fixes like prohibition, but requiring institutional permission to innovate, iterate and even compromise in assessment design.
The costs of compliance-based reform
A compliance-based reaction carries serious consequences, such as intensifying the workload of staff increasingly burdened with policing practices, interpreting ambiguous guidance and moderated contested outcomes. All of this adds to existing pressures on academic workload, without improving learning or outcomes for students. Despite there being tools to detect AI use, they are often unreliable and ethically complex, reinforcing mistrust and confusion between staff and students, rather than pedagogic clarity.
Where punitive policies are implemented, student engagement may be reduced or redirected towards risk avoidance, but this does little to support meaningful learning or support students in developing their own academic judgement. Given that graduates will enter a workforce in which generative AI is increasingly embedded, universities have a responsibility to support students in developing the critical, ethical and reflective judgement needed to use such tools well, rather than encouraging avoidance.
Finally, the risk to institutional reputation is heightened, increasing media coverage and regulator warnings are prompting calls for redesign rather than doubling down on detection regimes. However, to protect reputations, this needs to be meaningful, not simply substituting written assessments with oral presentations or similar surface level changes that ultimately reproduce the same underlying problems in a different format. Universities need to revisit the purpose, design and resourcing of assessment, rather than treating redesign as a technical fix for reputational anxiety.
Why this is really an institutional problem
Generative AI hasn’t created assessment problems – it has exposed them. Traditional formats such as essays, timed exams or standardised tasks have long been outdated and have never been aligned to key capabilities such as critical thinking, synthesis or ethical judgement. What AI amplifies is the misalignment between assessment and design and learning purpose.
The Australian Assessment Reform for the Age of Artificial Intelligence report highlights that existing challenges in assessment predate AI, but AI has sharpened the urgency to rethink practice rather than simply resisting technological change. Redesign strategies that integrate AI to support learning, depend less on the technology itself and more on how assessment purpose, workload and governance are aligned to enable meaningful use.
Toward trust-based reform
Universities must embrace trust-based assessment reform; this does not mean ignoring ethical concerns, but embedding purpose, capacity and alignment into assessment practices. Trust based reform entails:
- Assessment as learning: Designing tasks that require students to make decisions, justify reasoning, and demonstrate understanding in ways that cannot be outsourced to a tool
- Clarity and collaboration with students: Policies that make expectations clear and involve students in co-creating norms surrounding AI use
- Institutional support for staff: Recognising that redesigning assessment is skilled work that requires time, development and supportive governance
A provocation to sit with
Institutions are at a point where choices about assessment reform will define the future of higher education. If reform remains dominated by compliance and risk-aversion, institutions may succeed in preserving existing structures, but potentially at the cost of staff wellbeing, student trust and genuine educational purpose. Alternatively, if institutions leverage AI as a catalyst for rethinking why and how we assess, we can move towards systems that are coherent, sustainable and aligned with the true values and purpose of higher education.