The challenge of AI in Higher Education

In 2025, at a Higher Education conference in Melbourne I attended, several Victorian university Vice-Chancellors spoke glowingly of the potential for Artificial Intelligence (AI) on their campuses. A fair measure of their enthusiasm for AI related to the use of this technology to automate various administrative processes for their institutions, including ‘chatbots’ that will handle enquiries and guidance for students and staff. There was, however, little discussion on the substantial impact that AI is having on teaching and learning, as this technology rapidly infiltrates our schools and universities, magnifying the capacity for AI to automate nearly all aspects of education.

AI technology relies on advancements in computer capacity through sophisticated algorithms and access to vast troves of publicly available data to process information in order to simulate how we think human brains learn and produce knowledge. Underlying the approach are computational techniques that can handle large and complex data, particularly drawing links between data elements which enable predictions and outputs. A particular branch of AI, referred to as generative AI, can produce text, images, sound and other outputs in response to user prompts. However, the machine itself does not posses any insights, rather it produces outcomes based on statistical analysis, recognising patterns and data relationships, which is a simulation of human brain capacities but without any conscious overlay. AI development is currently fast-moving with vast financial investment, with the promise that AI will enable greater automation of a range of processes and gains in productivity. AI agents are available to help source information and produce outputs depending on the needs of the human user. Developments such as ‘agentic’ AI offer opportunities to automate complex series of tasks under computer control, requiring minimal human input.

The use of AI in high schools and universities is rife, as these tools not only offer the opportunity to enable research across the trove of publicly available information, but also offer help in writing part or all of assessment items. AI-generated assignments and essays can be difficult to detect, although the often superficiality in the output and writing styles can be obvious unless the user has sophisticated prompts. There are also AI approaches available to mask the computer-generated look and feel of writing work. AI is also there for teachers to use in increasingly sophisticated ways. AI will help with presentations, study plans and in the design of assessment items, as well as the marking and feedback on submitted work with reference to a rubric. If we are not there yet, it is possible to imagine computers setting the work and assessment, computers doing the assignment, and computers marking the assignment. The computers in this sequence all get a gold star. I know that we would normally assume greater integrity in these processes, but time-poor and social media-distracted humans confronted with deadlines may take the shortcuts that sever connections to actual human learning. Conversely, many universities are taking a ‘light’ or ‘responsible use’ approach to AI regulation, whilst also offering AI-based resources to students to assist in their studies, an approach that may seem necessary to remain competitive for future student recruitment. I am sure it has not occurred to university senior managers, but the more students are assisted to progress and pass their subjects, the greater improvements they will likely see in failure rates and student withdrawals.

There will also be voices that will, accurately, note that AI tools are here to stay, and we must adapt to the use of AI, perhaps to cordon off their involvement to search functions and assignment plans and outlines, rather than the assignment itself. AI will also likely have a major role in future jobs and so capacity to work with related tools and automation may be seen as desirable, if not necessary.

What we don’t have yet is a systematic and rigorous approach to the role of AI in learning and teaching. Educational guidelines will emphasise the responsible use of AI but without hard detail on what is acceptable or not – what are the thresholds? For teachers, how much of their work can be substituted with AI generated materials and approaches? How can we institute practices that will ensure our university graduates have the credited discipline-specific knowledge and skills of their qualification, as well as the capacity for critical analysis? Will this be mostly redundant for a range of fields where machines will do a good-enough job, replacing the human element? We have recently seen that AI machines can reliably undertake significant work in software coding, and can also replace the work done by specific software packages, which puts at risk entire fields of work and software-based industries. Undoubtedly, AI companies will bring this focus to other industries and the broad field of education is likely to be subject to wholesale disruption. We have seen previous alarms of digital disruption in higher education, notably in the proliferation of online learning, Massive Open Online Courses, and other short course approaches that have undermined degree-based approaches. These didn’t bring down universities but they did open up access to new and extensive learning opportunities, calling into account the value of ‘credentialed’ learning via traditional degrees.

The unique genius of the human brain is its capacity for learning as well as its plasticity, written into the formation and reformation of synaptic pathways in response to stimulation. This process doesn’t stop in early adulthood, with lifelong learning and cognitive stimulation adding to brain function, as well resilience, known as cognitive reserve, to neurodegenerative conditions such as Alzheimer’s disease. The human brain is also supremely good, and sometimes a bit lazy, in its ability to project through tools, helping to reduce the burden of work, and leading to a variety of gains. Hence, we are drawn to power tools over screwdrivers and saws, computers and keyboards over pens and paper, cars over walking, social media over slower forms of information gathering and personal communication. The real danger around AI is our brain’s capacity to happily substitute computer processes for real cognitive work, hence, leading to brain adaptations that will reflect in poor skills in deep learning, critical analysis and communication. We may find future brains will be unable to undertake complex cognitive work without the substantial assistance of the machines.

If this acceleration in AI technology is the start of what has been described as the fourth Industrial Revolution, there is little to be gained from resistance to our further integration with the machines, but we must also consider a re-engineering of education to ensure we remain masters of the technology, and that our brains are not down-sized in their capacity for cognitive work. There is already some research emerging showing how engagement with AI leads to ‘cognitive offloading’ amongst students, a dangerous outcome for our overall brain health. It will be essential to purposively address the challenges as well as the opportunities that AI-based technology will bring in education as well as a range of other fields. Alternatively, it may be that we will soon be welcoming the agentic University, trained to package up the world’s knowledge in educational bites and delivered to any of our digital devices as required, but that also ensures that machines will be our constant tool.