Even as an institution who probably has all the necessary educational policies regarding GenAI in education in place: policy in itself is not enough. People are struggling. It’s in trying to make sense of the on-the-ground experiences that I found a powerful lens in a paper by Corbin et al. (2025), which frames AI in assessment not as a puzzle to be solved, but as a “wicked problem” to be navigated.

Why ‘Wicked’? A reflection on our shared reality

The term “wicked problem” might be academic, but reading the Corbin et al. paper was a lightbulb moment. As one reaction to the article perfectly put it, “This is comforting. And discomforting” – a sentiment that instantly resonated. It’s comforting because it gives a name to the complex, often contradictory situations many of us encounter every day in educational institutions. It’s discomforting because it confirms there is no easy fix. The article’s hallmarks of a wicked problem aren’t just theoretical; they are a direct reflection of the conversations happening in staff rooms and development meetings everywhere:

  • Confusion, division, and a lack of consensus: I see how educators experience this problem as confusing and intangible. Many admit to being at a loss and frustrated, unsure how or if to react. There is no shared understanding of what needs to be addressed; people disagree on the goals of assessment, the appropriate formats, the heart of the problem and the resources required. This division makes a unified approach impossible and reinforces a sense of helplessness.
  • Every action has a consequence: The article highlights a familiar cycle: educators endlessly tweak assessments to add value or security, yet never feel the design is “good enough.” The stopping point is usually a deadline, not a sense of completion. Critically, every one of these refinements directly impacts students, meaning there are no risk-free experiments, which only raises the stakes.
  • AI as a symptom of deeper systemic issues: The challenges surrounding GenAI are often a symptom of broader, structural problems: student groups that are too large, not enough time for meaningful evaluation, and a focus on final results over the learning process. GenAI has not caused these tensions, but it has certainly exposed them. The problem is rooted deeper within the educational system itself, which makes it all the harder to distance from these traditions.

Shifting our approach: from solution to navigation

Across institutions, there’s often a hope for a top-down mandate—a clear policy that will just tell everyone what to do. But from a learning design perspective, we know this is a fallacy. Teaching is a design science; it relies on patterns adapted to unique contexts.

This is where our role as educational developers becomes so crucial. Our task isn’t to deliver a silver-bullet solution, but to empower our academic colleagues to become confident designers in this new reality. Our conversations need to shift from finding a single “answer” to fostering professional autonomy. The goal, in my opinion, is to help our colleagues move from a state of paralysis to empowered action. Based on emerging practices, this involves:

  1. Expanding the toolkit: We can start by illuminating the full spectrum of assessment possibilities. By showcasing a diverse range of methods and approaches, we open up the choices available, allowing educators to select, build and contextualise assessments that are right for their discipline, their students and their pedagogical goals and values.
  2. Championing a design mindset: We also need to champion a framework that encourages reflection, iteration and adjustment. What works in a philosophy seminar of 20 won’t work in a first-year engineering lecture of 300. We can help create the institutional permission for staff to experiment, accepting that some attempts will fail, that compromises are inevitable and understanding that this iterative process is a core part of professional educational design.
  3. Acknowledging the human: Finally, it’s vital to acknowledge the messy human elements that make up education and work. In these conversations, it’s clear this shift is unsettling. It challenges professional identities and can create a sense of loss for established practices. Part of our job is to acknowledge this and support colleagues through that feeling of uncertainty, helping them (re)find their footing.

Imperfect action is better than none

To frame a challenge as a ‘wicked problem’ is not to admit defeat, as the article stresses, but to empower ourselves with realism. It releases us from the need for a perfect, final solution and instead focuses us on finding better responses. And the way I see it: the most damaging response is inaction born of fear or the search for a flawless strategy.

Perhaps our most important job right now is to help our colleagues take those first, imperfect steps. Let’s provide the support, the options and the space they need to navigate this new terrain with professional skill and renewed confidence.

Based on discussions and thoughts around: Corbin, T., Bearman, M., Boud, D., & Dawson, P. (2025). The wicked problem of AI and assessment. Assessment & Evaluation in Higher Education, 0(0), 1-17. https://doi.org/10.1080/02602938.2025.2553340