The major debate currently surrounding artificial intelligence in higher education asks whether AI should be policed. But in focusing so heavily on control, we overlook the culture of shame this creates. The result? Instead of learning to use AI effectively, students are just finding ways to hide their use of it.
So, if higher education is to remain relevant, we need to move beyond this binary thinking. We need to stop arguing about whether AI belongs in the classroom. It’s already here and is demanding that we challenge some of our entrenched thinking and practices.
Take the fact that higher education has historically rewarded not only what students know, but how cleanly they present it such as correct grammar and structure, as well as academic fluency . The problem is that many of these feature that were once used as proxies for academic ability can now be automated almost entirely by AI . And often, it does so in ways that feel superficially polished and ultimately hollow, what many label as AI “slop.”
In this context, the value of learning has to shift. We need to move away from being output-driven, focusing more on the process and showing one’s thinking. Because true learning happens when students grapple with ideas and sit in the discomfort of ‘not knowing’ or, rather, ‘on the way to knowing’. This kind of productive struggle cannot be outsourced. It requires engagement and, importantly, imperfection.
And yet, higher education has historically done the opposite. We’ve rewarded perfection and penalised uncertainty. Students have been conditioned to only speak when they are sure, to present only what is polished. AI simply accelerates that tendency.
So what if we actively designed for the kind of conditions that make learning possible?
Now that machines can generate fluent, coherent, and often convincing text, what becomes valuable is a greater presence of one’s humanity. One area that cannot be easily replicated by AI is our embodied, lived (and naturally imperfect) experience. A student’s ability to take knowledge and apply it to their own context, interpreting and embedding it in their reality, helps to give meaning to their learning, requiring their own insights and experience. Increasingly, we’re seeing this play out in society too, where work that is recognisably human, carrying an authentic voice and perspective, is ascribed greater value than something that feels like generated content.
This requires a fundamental rethink of what we ask students to produce. If a submission is so polished that the student disappears from it, then we have lost something essential.
A new role for educators
If AI can provide information faster and more efficiently than any lecturer, then the role of the educator cannot remain that of sole knowledge custodians. We need to rather see ourselves as facilitators of learning, which in turn requires us to create conditions for the struggle that enables understanding.
This means creating friction in the learning process: asking harder, even uncomfortable questions that require application, reflection, and critique. It means designing spaces where students engage and challenge each other to think beyond the surface-level response. The process may be less tidy, but if it’s more engaging and memorable, it’s ultimately more conducive to meaningful learning.
It also means working with AI, not against it, allowing it to be the most efficient tool in the room, but not mistaking that for wisdom Students should learn how to harness its speed and accuracy, while recognising its limits. It cannot replicate judgement, context, or the distinctly human qualities that shape real understanding — humour, emotion, and even the insight that emerges through error. This isn’t a comfortable shift and requires that educators rethink their position, pedagogy, and, in many cases, identity.
Redesigning assessment for an AI world
If the purpose of learning is changing, then assessment must change too.
The current system, built around traditional assignments and examinations, needs to be reimagined.
At Eduvos, we’ve done this by embracing a balance. There is still a place for academic integrity measures, including detection tools, but these are used cautiously, as indicators rather than evidence. They serve as flags that may prompt further review, not as definitive judgements of misconduct. Our focus is not on punishment, but on engagement. Where concerns arise, our starting point is to find opportunities to better guide students on productive use of AI.
More importantly, learning and assessment activities must actively integrate AI.We ask our students, for instance, to actively use AI in their work, such as building tools or working with agentic AI to demonstrate how they engage with the technology to apply that knowledge. This shift is resource-intensive and disrupts long-established systems. Without it, however, we risk preparing students for a world that no longer exists.
Learning as shared responsibility
There will always be students who may seek the path of least resistance. And it’s become a lot easier to do just that. This means we need to think of education as more of a shared responsibility between the institution and the student. In this conception, students must choose to engage, show up, struggle, and resist the temptation to outsource their thinking entirely.
Institutions, in turn, must equip them for an AI-enabled world. This means building AI literacy, demystifying tools, and providing clear, practical guidance on how to use them responsibly.
It also means rethinking what we reward. If we continue to prioritise perfection, we will continue to incentivise AI over authenticity.
In many ways, AI is moving higher education closer to its original purpose. At its core, education was never meant to be transactional. It was meant to be dialogical, focused on engagement, conversation, and the co-creation of knowledge.
Today’s students are asking for exactly this. They want authenticity and relevance. And they want learning that connects to the real world and their own lived experiences.
AI, paradoxically, may be the catalyst that brings us back to that vision but only if we’re willing to move beyond fear and policing, and towards a more integrated, collaborative approach.
The future of higher education will not be defined by whether we allow AI into the classroom. It will come down to whether we are able to redesign learning so that, even in the presence of AI, the most valuable thing a student brings remains unmistakably their own.
