top of page

In the age of AI, universities must manufacture the cognitive struggle

  • May 28
  • 4 min read
A new study reveals that over 90% of students now use AI in their education, posing a risk of cognitive decline. For universities, their value now lies in encouraging active mental exertion.

The following insights are drawn from a May 2026 roundtable focused on the fundamental challenges facing education systems as they prepare learners for an uncertain, AI-driven future. The provocateur for this roundtable was Melissa Loble, Chief Academic Officer,

Instructure. Drawing on decades of experience as both an educator and an executive, Melissa spearheaded the integration of academic rigour and customer-centric strategy across the Instructure learning ecosystem (Canvas LMS), ensuring the educator experience remains central to digital innovation.



AI use is nearly universal among undergraduates. Recent data highlights the scale of adoption: a 2026 Higher Education Policy Institute (HEPI) study reveals that 92% of students use AI in their education, with 88% utilising it during assessments. And according to the Digital Education Council, 57% of students employ generative tools at least weekly, and over 30% interact with them daily (despite many expressing mixed feelings about AI more broadly). 


This is a decentralised, "bring-your-own-AI" phenomenon. Students are independently experimenting with AI, faster than universities can introduce approved toolkits or integrate it into curricula en masse, creating a wide adoption gap between learners and a far slower-moving faculty. While educators and leadership teams are also actively trying to engage AI, they simply are not matching the pace or volume of the student body.


This discrepancy forces a fundamental question (one which is the focus of ongoing academic research): is AI usage helping learners build more meaningful, durable skills for a fast-changing workforce, or is it fundamentally hampering our long-term cognitive abilities? 



The risk of cognitive slide


The baseline concern raised by roundtable participants is that AI can too easily reduce the cognitive effort required in the learning experience. Recent graduates and students observe that the introduction of AI has effectively "kicked out the need for effort" from traditional academic milestones. When the messy, multi-week process of researching, drafting and refining an essay (for example) is entirely offloaded to a quick prompt, the cognitive skillset that the assignment aimed to build isn’t developed.


Just like physical fitness, cognitive development requires a baseline of friction to build and maintain neural pathways. If younger generations are handed an "AI bike" before they learn how to wobble, fall and find their balance, some fundamental cognitive muscles may never develop at all. 


While research is still ongoing, emerging studies indicate that over-reliance on AI tools can diminish active brain pathways and reduce cognitive resilience. Because new generations will grow up entirely immersed in automated environments, their brains will inevitably be wired differently than those of their predecessors. Without continuous scientific tracking, we cannot yet know the full impact of this automated shift on long-term human intelligence.



Encouraging effort in the AI era


As traditional knowledge structures dissolve, universities face the rapid commoditisation of the "right answer." Industry leaders at the roundtable noted that the half-life of raw technical knowledge is now measured in months. Consequently, employers are no longer hiring individuals for what they have already learned; they are actively seeking continuous learners who can adapt, collaborate and co-create alongside technology.


Real learning cannot be synthesised or bypassed. It is fundamentally an endothermic biochemical reaction, requiring a deliberate input of human energy to change states in the brain and thereby encode lessons learned. AI can act as an exceptionally fast catalyst, but the reaction itself cannot occur unless the underlying human effort remains intact.


Acknowledging this, educators can add real value by designing friction-filled, experiential and multi-system challenges that demand active mental exertion. These learning experiences force students to think on the spot, build communication skills and tap into their own knowledge and instincts, thereby developing the exact skills employers are looking for.



Assessing beyond knowledge and skills


Introducing more experiential, social learning exposes a common institutional bottleneck: grading final outcomes, such as essays and exam papers, is simple and scalable, but evaluating the internal process of human thinking poses more of a challenge. One participant suggested dividing human competence into three dimensions: knowledge, skills and 'behaviour'. What they meant by ‘behaviour’ was the cognitive processes and habits that underlie how learners actually apply knowledge and skills in authentic situations.


While higher education has historically excelled at evaluating the first (knowledge), and across the last couple of decades has become much better at the second (skills or competencies), it has spent far less time designing scalable ways to assess the third (cognitive processes and habits). But learning to measure and reward these dynamic patterns is essential for any university looking to stay relevant.


Universities must design authentic situations in which students enact judgement, collaboration, and metacognitive habits — and develop evaluation methods that observe and feed back on those enactments rather than only the final artefact. This must sit alongside developing students' capabilities in ethical prompt engineering, and cultivating a genuine passion for learning.  To support this transformation, universities must urgently upskill their own faculty, providing the training and structural incentives required to integrate meaningful AI orchestration into every curriculum.


Cultivating metacognitive resilience


To safeguard against cognitive decline, educators and employers need to align their practices with what neuroscience is telling us. Building neuroplasticity — which underpins continuous learning, innovation and resilience — requires iterative knowledge retrieval and skill practice, and a continued willingness to cognitively struggle. When we rely on AI to answer every question rather than expand our knowledge base, we undermine our own memory; when we rely on it to write our essays rather than challenge us with brainstorms and critiques, we minimise the cognitive struggle that triggers true learning and growth.


For universities, the ultimate goal should be fostering metacognitive habits — helping students to become conscious of how they learn and cultivating lifelong habits that support continuous learning. When experiential learning forces a student to step back and reflect on their interactions with automated tools, they build a robust, change-ready resilience that no algorithm can replicate or replace. The human remains the sculptor, treating AI strictly as the chisel.




Thank you also to everyone who attended this roundtable. We look forward to continuing this urgent dialogue as we actively work with partners across education and industry to bridge the gap between the classroom and the world.



bottom of page