What universities are getting wrong about teaching in the age of AI
Skills training alone won't prepare graduates for a world where AI is doing the technical work. The real fix lies in how universities teach, not what
Published 29 June 2026
It's an understatement that educators worry about students using AI to offload the cognitive struggle that is critical for learning. That worry is well founded.
Employers who must train the next generation in their professions are wrestling with the same concern.
Traditionally, early career professionals spent their first years in the workforce doing technical work, slowly building the mental models their expertise will be built on.
But all that technical work will soon be done by machines.
Educators and employers share the same fear – machines do the work, supervised by professionals who lack the expertise to tell whether the output is competent.
Our research, as well as our years inside engineering classrooms and fostering industry projects, suggests the likely response will not get at the core issue. That response, more contextualised project-based work injected into the traditional syllabus, misses the point.
We need to focus on how students think about their learning.
Our recent research looked at how a subject that paired multidisciplinary student teams with industry sponsors to solve real-world strategic challenges, increased graduates' employability.
The subject aimed to do something unusual. Instead of just building skills, it taught students to think like working professionals and to learn the skills in that context.
We asked former students, six months to six years after graduation, about their experiences.
The biggest driver of student employability was how they came to see themselves, not the skills they learned.
Those who started to think like professionals picked up the skills they needed and used them. Those who kept thinking like students did not.
We measured four aspects of that thinking shift.
Did students come to believe ability can be grown rather than being fixed? Did they start to see professionalism as creating value for others?
Did they start to take personal responsibility for their own careers? And did they come to see excellence as something a team produces, not a lone individual?
The more a student moved towards these four beliefs while taking the subject, the more they were using the skills they had been taught in their current jobs, and the better they scored on three measures of employability.
Those employability measures were clarity about the job they wanted, the ability to land it and the ability to do it well.
Unlike traditional learning, in which students acquire knowledge or skills within a fixed worldview, this is an example of transformational learning – learning in which students change their worldviews.
Transformational learning mattered because students who started thinking like professionals became genuinely motivated to learn these professional skills. They wanted to know when and how to use them.
Learning professional skills was no longer just a way to earn high marks. It became a challenge they found interesting and cared about, regardless of their grade.
Our findings speak directly to a risk we now face.
The subject helped students become employable by doing two things at once: shifting how they saw themselves, which created real motivation, and teaching them professional skills consistent with that motivation.
The same is true for anyone trying to educate students and young professionals in the age of AI. Students have to actually want to learn what their instructors are there to teach.
If education stays built on the idea that a student's job is to jump through hoops set by their lecturer, students will always be tempted to hand the hard mental work to AI.
The danger here is that educators and students end up in a never-ending arms race between hard-work offloading and offloading-detection. On the other hand, students who genuinely want to learn will focus far more on their learning and far less on their marks.
Likewise, young professionals will always have the machine next to them, tempting them to take the easy, quick route to the correct answer.
Only genuinely motivated graduates will reliably embrace the hard work required to become experts. They need to know why they should do the hard work and value that ‘why’.
We have only just started to grasp the implication, that in the age of AI, educators may need to build motivation before they build skills.
But we have ideas.
Firstly, education will have to include transformational learning subjects that bring students’ world views into line with their chosen profession.
This is not as Orwellian as it seems.
The principal goal of medical schools, engineering schools and fine art departments has always been to teach their students to think like doctors, engineers and art historians.
This process just needs to become explicit.
That shift will lead students to value the experiences essential to their growth, even the boring or difficult ones.
Health & Medicine
As AI gets smarter, are we getting dumber?
There are many other ways to reach students' motivation, whether you're aiming for this kind of transformation or working toward something more modest.
Students are more likely to be motivated when they have greater ownership over what they learn, when assessments push them to teach each other, and when assessments deliver other things they value, like closer relationships with their peers.
But these changes may also require deep shifts in the craft of education.
The dynamic between instructor and student, for one, may need to change. Educators may need to look less like authority figures who own the learning and more like partners who help it along.
While knowledge and skills education remain necessary, they were never sufficient, and with AI will be less so.
The future of professional education lies in experiences that build and exploit true motivation, by drawing on other internal drivers and reshaping how students see themselves.