Generative artificial intelligence, which can produce outputs that map closely to many of our assessments, can seduce students into thinking they have mastered learning outcomes without achieving mastery.
Rather than allowing learners to succumb to the “illusion of learning”, in the words of one of our students, we believe in fully embracing the effort involved in taking on tasks beyond our comfort zone. We also argue for an expansive definition of “productive struggle”, one that moves beyond cognitive aspects to include social and relational dimensions. Challenges that require productive struggle are vital for supporting our students to create a better world and crucial for preparing the next generation of society’s leaders.
Although originally developed for secondary-school mathematics education, Hiroko Warshauer’s idea of productive struggle has become central to discussions of GenAI. When supported by teachers or peers, students who work through challenges demonstrate deeper understanding and stronger knowledge retention than those who disengage. GenAI tools, however, have the potential to minimise or even remove struggle altogether, which disrupts this process. They enable learners to circumvent traditional strategies, such as breaking tasks like essay writing into stages, that are designed to support students through moments of difficulty.
- Confusion diffusion: nine ways to prepare students to work with uncertainty
- Struggle, strength and growth: preparing students for the changing future
- Learning challenge and difficulty: calm seas do not make skilful sailors
This can lead to cognitive offloading and, more critically, cognitive surrender, in which users defer entirely to GenAI, accepting its outputs without engaging their own judgement or critical evaluation, as Steven D. Shaw and Gideon Nave from the Wharton School have found. Using AIs prematurely fuels the exact illusion of learning our student identified.
If students use GenAI in the scaffolded process, does it simply offload a cognitive task – in a similar fashion to using a calculator – or does it impede the intended learning? We are still working to understand how much practice is required to develop the experience needed for productive collaboration with AI tools.
Using challenge to deepen understanding
Learning is both a cognitive and relational activity, modelling the kinds of community relationships we value. As such, productive struggle is vital for relational and social learning. At our institution, for example, we begin our meetings by acknowledging the land and our relationships to all our relations – human and more-than-human – in this place.
Brazilian educator and philosopher Paulo Freire challenged educators to teach students to read not just the word but the world. GenAI excels at reading the word, ingesting vast amounts of text to predict what comes next. Yet, Freire reminds us, education is not simply about absorbing knowledge but transforming it, layering in our own understanding and lived experiences, and critically engaging with the social, cultural and political dimensions embedded within it. This responsibility requires us to actively seek out voices and ways of knowing that are often excluded from AI outputs, such as texts written in languages other than English and knowledge systems carried through oral traditions.
So, how can we leverage productive struggle to encourage our students to read the world as well as the word?
We can create conditions for productive struggle through assessment design. Often, students struggle with lack of definition. So, there is value in ambiguity or open-ended tasks that require our students to spend time on deciding how they want to approach an assignment. For example, we can assign a task relevant to a community service student’s placement. Students may need to think about what clients a partner organisation supports and associated socio-political issues. They will be required to define a question, develop an argument, conduct research and prepare a paper or presentation for a specific audience (such as their instructor, a partner organisation, clients or classmates). Once they see the relevance of their work and the impact it has had, they demonstrate significant pride.
Beyond assessment, we must also consider how structures within our institutions can support productive struggle. The story of an engineering student helps illustrate this. After facing an academic setback, the student participated in our Fresh Start programme to re-enter university and pivot towards the interdisciplinary field of science, technology and society. This transition allowed her to merge her technical foundation with her experience as a food bank volunteer to analyse complex social systems. Now an award-winning undergraduate researcher, she employs institutional ethnography to grapple with the paradox that campus food waste occurs alongside acute student food insecurity.
On our campuses – in person and online – we need to make space for productive struggle, even when the policies or decisions of administrators are the focus of criticism (for example, about food waste). As we reimagine learning structure and assessment in higher education, we must view such struggle as a sign that we are doing our job. We can reframe this tension as an opportunity to co-create with our diverse communities and students.
As humans, we have a responsibility not only to predict what comes next but also to decide what should come next. When we surrender this responsibility to AI, we diminish not only our individual ability to learn but also our collective capacity to address complex societal challenges.
Karsten Mundel is vice-provost (learning initiatives) and Olivia Murray is an instructional designer (multifaceted evaluation of teaching and learning) in the Centre for Teaching and Learning, both at the University of Alberta, Canada.
If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.
comment