Opinion
How can universities make good decisions when technology changes faster than governance?
I was asked recently about Macquarie’s position on wearable technologies.
The question had been prompted by an article a colleague had read considering whether universities should prohibit AI-enabled wearable devices during assessment tasks.
It is an important question, and one that will inevitably become more common as smart glasses, watches, earbuds, rings, and other devices become increasingly capable.
I did not immediately have an answer.
My first instinct was to consider the devices themselves, their capabilities, the circumstances in which their use might be permitted, and when they might need to be excluded from an assessment environment. But it quickly became apparent that rules of this kind would risk being overtaken by technological development almost as soon as they had been written. I realised I was probably asking the wrong question.
The more fundamental question is not what Macquarie, or universities more broadly, should do about wearable technologies. It is how we should exercise sound educational judgement when technological change has become a permanent condition of academic life. It is easy to focus on the technology.
What distinguishes the present moment from earlier periods of disruption, however, is less the fact of change than its pace. It is this pace that requires principles and reasoned judgements that endure without generating an ever-expanding number of rules.
On the same day, I was also asked whether we should adopt a particular AI detection technology. Although apparently distinct, the two questions arise from the same underlying problem.
Universities, and university leaders, have long been accustomed to governing through periods of relative technological stability. New technologies emerged at a pace that allowed institutions to observe, evaluate, consult, and respond.
Artificial intelligence has altered that relationship. Technological capability is now evolving so quickly that institutional governance can barely keep pace. By the time a policy has been debated, approved, communicated, and implemented, the technology to which it responds may already have changed. The question about wearables will soon be overtaken by questions about the next generation of intelligent technologies, and then by others beyond them.
The rapid advance of AI-enabled technology creates a strong and understandable temptation. Faced with accelerating uncertainty, institutions may seek assurance through ever more detailed rules. Yet, every new technology cannot require another policy, prohibition, clarification, and set of exceptions. If it does, our governance systems will become increasingly specific, cumbersome, and ultimately inert.
There is something of a paradox here. The more quickly technology evolves, the less useful technology-specific policy becomes.
This is not because governance becomes less important. It is because the object of governance needs to change. Where technological change is episodic, it may be possible to govern particular technologies; but where technological change is continuous, institutions should govern through enduring principles.
Wearable technologies, AI detection software, and the next generation of intelligent systems are often presented as separate policy problems. In reality, they are manifestations of the same underlying challenge.
Each invites us to ask what rule should apply to a particular technology. Yet, rules framed around particular technologies inevitably age at the same pace as the technologies themselves. Principles, however, endure because they begin with a different question.
Rather than asking whether a particular technology should be prohibited, we might first ask what evidence of learning an assessment is intended to establish. This is not a new insight, and it has featured prominently in discussions of AI and assessment. Nonetheless, it remains the most useful starting point. Where independent reasoning, personal competence, or unaided performance forms part of the intended learning, restrictions on technology may be entirely justified. Their justification lies not in the existence of artificial intelligence, but in the educational purpose of the assessment.
Conversely, where graduates will need to exercise professional judgement while working alongside increasingly capable AI systems, assessment should reflect that reality. Educational purpose should determine assessment conditions, not the other way around.
The same reasoning applies to the continuing debate about AI detection software. Much of that debate has focused on technical reliability. Detection tools produce false positives, struggle with edited text, and become less reliable as generative models improve. These concerns are well placed. Even if detection technologies became highly accurate, I would remain cautious about making them central to institutional practice.
My principal concern is educational. Detection technologies risk encouraging universities to preserve forms of assessment that may already have been losing educational value before generative AI emerged. They invite us to ask whether students have used AI when the more fundamental question is whether our assessments continue to provide valid evidence of learning. Artificial intelligence is not responsible for every weakness in assessment. In many instances, it has simply drawn attention to weaknesses that already existed.
None of this diminishes the challenges confronting academic colleagues. They are considerable. Staff are being asked to navigate profound technological change while maintaining academic standards, supporting students, and redesigning elements of their practice. Their responses are understandably diverse. Some are enthusiastic, others cautious, and many remain uncertain about the pace and direction of change. Students occupy much the same spectrum. Academic leaders are expected to provide confidence at precisely the moment when certainty is least available.
The instinct to provide increasingly detailed institutional direction is therefore understandable. Yet, universities cannot remove uncertainty by multiplying rules. What institutions can provide is greater clarity, consistency, and confidence over time, grounded in educational purposes that do not change whenever the technology does. This requires us to recover and renew an enduring feature of universities.
Universities have never depended primarily upon exhaustive regulation. They have relied upon professional judgement exercised within a framework of shared principles. Foundational practices such as academic freedom, peer review, disciplinary standards, curriculum design, and assessment all depend upon the capacity of academic communities to exercise informed judgement responsibly.
Rules have always played an important role, but they have never been a substitute for judgement itself.
Artificial intelligence does not diminish the importance of that judgement. It makes its careful exercise more important than ever.
This may be the defining governance challenge of the coming decade. The greatest risk facing universities is not that artificial intelligence will outpace our policies. It almost certainly will. The greater risk is that we respond by replacing academic judgement with an ever-growing catalogue of rules. Such an approach may create the appearance of certainty, yet it is unlikely to produce better educational decisions.
This does not absolve institutions of the responsibility to provide practical assistance. Academic colleagues need timely guidance as new questions arise, clear assessment parameters, opportunities to develop their practice, and confidence that they will be supported in making difficult judgements.
Principle-based governance cannot mean leaving each academic to navigate technological change alone. The role of the institution is to provide a coherent framework and meaningful support without seeking to prescribe an answer for every circumstance.
There is a related challenge in assuring an increasingly sceptical public that academic standards will be preserved in an AI-enabled era that deserves consideration in its own right. Public assurance cannot rest solely upon the visibility of controls. It will increasingly depend upon universities being able to explain what students have learned, how that learning has been demonstrated, and why the judgements underpinning assessment remain worthy of confidence. We have fallen short in our capacity to assure, and that is work we also need to do.
Universities have encountered profound change many times before. The expansion from elite to mass higher education transformed who universities served and required institutions to reconsider how their purposes could be realised at vastly greater scale. Authoritarian regimes tested whether commitments to academic freedom, scholarly integrity, and institutional autonomy were merely convenient practices or defining principles.
The arrival of the internet fundamentally altered the relationship between knowledge, information, and expertise.
Universities did not always navigate these changes well. Nor did first principles provide simple answers. The most enduring adaptations, however, came from distinguishing between what had to remain constant and what had to change. Principles provided continuity. Judgement enabled adaptation.
The question, then, is not simply whether universities should ban AI-enabled wearables, nor whether a particular detection technology should be adopted. It is whether, in an age when technology changes faster than governance, we have sufficient confidence in our educational principles, and in the professional judgement of our academic communities, to govern wisely without believing that every new technology requires another rule.
Rorden Wilkinson is Deputy Vice-Chancellor (Academic) at Macquarie University.