Transforming Education: A Systems Thinking Approach to Assessment in the AI Era

AI is transforming knowledge work and fueling discussions around academic integrity and assessment security in higher education. While these concerns are legitimate, strategies being proposed often focus on the symptoms rather than addressing the underlying issues.

What if the real challenge isn’t AI itself, but how we approach and define assessment?

Rather than viewing AI as a threat, it can serve as a timely lens to reexamine conventional assessment models often structured as judgment events rather than as part of the ongoing learning process. A point well illustrated by my colleague Jason M. Lodge recently using the e-Bike analogy.

As the World Economic Forum White Paper highlighted in 2023, this outdated “mental model has dominated higher education for centuries, but AI is finally forcing us to confront its limitations.

The need for change isn’t just theoretical—it directly impacts students. In my recent AI session with a group of medical students at The University of Queensland, one student’s insight stood out: We’re left out of these [AI impact on future work; AI work-readiness; AI and Work-Ready Assessment] conversations, yet we’re the ones navigating rigid academic requirements while feeling pressure to prepare for AI-driven clinical care. What truly resonated with me was their concluding observation: We need assessment approaches that acknowledge this reality. This highlighted a critical gap— while student partnership is valued in assessment for inclusion practices, students have not been as active partners in redesigning education to reflect the real-world capabilities needed in this AI-era - in part due to the complexity of navigating a rapidly changing world. A systems thinking approach might be helpful to support the change needed in higher education curriculum and assessment design now.

A More Holistic Solution: Embracing a Systems Thinking Approach

The complex nature of universities can often focus effort on short-term, or “low-hanging” fixes like proctored exams, or AI to support conventional assessment paradigms like assessment item development. While useful, these reactive measures don’t address the deeper issue.

A systems thinking approach recognises education as an interconnected complex network, where change in one area (like assessment) has a ripple effect on every other area of teaching and learning, and the institution. It helps us understand both the visible (e.g., policies, funding) and invisible (e.g., mental models, power dynamics, relationships) factors that shape our system.

Systems thinking shows us that lasting change requires collaboration across the entire educational ecosystem – no single program, organisational unit, or individual can achieve this alone. Complex change is not linear but dynamic, requiring us to adapt as the system evolves, thus ultimately achieving meaningful and sustained change. Supporting this notion, Fullan, and Dhukaram and colleagues proposed that nurturing a team of systems thinkers is critical to sustainability for large scale change in education.

As an example to illustrate a systems thinking approach, let's consider how the interconnected flow-on consequences of implementing AI-driven feedback could fundamentally reshape the assessment and teaching ecosystem. As students receive continuous insights on their development, they engage in deeper learning, requiring faculty to adopt more adaptive teaching and support strategies. This shift also may influence assessment structures, moving away from high-stakes testing towards formative, data-informed learning pathways. In turn, universities must rethink policies, technology infrastructure, faculty training, professional support structures, curriculum design and assessment policies to fully integrate AI as a supportive ally to student learning.

Evolving Faculty Roles and Development

The impact of AI-driven change highlights the evolving nature of faculty roles, creating opportunities for new approaches to teaching and assessment. AI can handle routine feedback and analytics thus allowing faculty to shift from assessors to facilitators of deeper learning and critical thinking. Less time can now be spent on grading, and more time can focus on coaching students, interpreting AI-driven insights, and guiding critical thinking and professional identity formation. This means a greater emphasis on mentorship, higher-order reasoning, and ethical considerations—areas where human expertise is irreplaceable. González-Calatayud's work showed that faculty development is the significant enabler to rethinking assessment. Faculty might require support to effectively integrate AI into education practices to ensure that technology effectively enhances learning rather than simply used to make outdated practices more efficient, or to replace meaningful human interactions.

Ultimately, a systems-based AI integration could encourage a cultural shift in higher education—one that redefines faculty roles and responsibilities, assessment design, and student engagement. By embracing this transformation, universities can move beyond AI-proofing assessments to building an education ecosystem that prioritises continuous learning, human-centered skills, and adaptability.

Rediscovering the True Value of University Education

To move beyond reactive strategies, universities must explore what truly sets them apart in an AI-driven world. Higher education’s greatest value extends beyond delivering information—AI can increasingly handle that—but in cultivating expert-guided learning through engagement with discipline scholars, industry leaders, and mentors who model expert reasoning and provide personalised guidance.

As Megan Lily Executive Director, Ai Group CET, notes: Today’s fast-paced technology landscape and rapidly evolving skill needs demand closer collaboration between universities and industry—not just in research, but in learning, content delivery, micro-credentials, and work-integrated projects.

To remain relevant, universities can embrace their role as hubs of critical thinking, innovation, and ethical reflection. This involves an ecosystem that fosters:

• Collaborative spaces where students, faculty, industry and AI work together to co-design, co-develop, and critically explore new ideas.

• Adaptive skill development that equips both students and faculty for rapidly changing environments and emerging technologies.

• Ethical and social growth by nuturing wisdom, judgment, adaptive leadership and human-centered decision-making that goes beyond technical skills.

• Alignment with professional practices to ensure graduates are prepared for workplaces where AI collaboration is integral, while also developing the uniquely human-centred skills increasingly valued by employers and communities.

AI is an essential part of this evolving ecosystem, offering real-time, personalised insights that support student development and targeted improvements. But while AI plays a crucial role, human mentorship and contextual understanding and the ability to foster deep critical thinking remain indispensable.

Instead of viewing AI as an inevitable challenge, universities could reaffirm what sets them apart and integrate AI as an ally to learning—amplifying faculty expertise, streamlining repetitive tasks, and enabling educators to focus on what truly matters: guiding students toward meaningful learning and professional identity development.

Ensuring Equity in AI-Enhanced Assessment Systems

A systems approach to AI can transform assessments, but it also raises a crucial question: Who might be left behind? As we redesign assessments, we must ensure that new models do not exacerbate existing inequities. In discussions with students, I have observed that we often assume a level of digital fluency that not all have developed. Students from diverse socioeconomic backgrounds experience varying access to AI tools and digital literacy skills. For example, rural students may face infrastructure limitations that urban peers do not. Additionally, some students are hesitant to experiment with AI for fear of unintended consequences, such as compromising academic integrity.

Breaking the Assessment as Event Paradigm

Ensuring fairness is just one part of the equation. We can fundamentally rethink assessment itself. Moving beyond the ‘assessment as event’ paradigm is a critical shift in educational design thinking – a AI-driven curriculum design threshold concept, if you like. High-stakes exams often fail to reflect authentic learning or professional identity development – critical aspects of university learning.

Faculty readiness to harness this paradigm varies widely across disciplines and individuals. While some might enthusiastically embrace AI-enhanced assessment approaches, others may have valid discipline-specific concerns or face steeper learning curves. The challenge for medicine for example, is very different for journalism or for engineering.

Building the Educational Ecosystem of Tomorrow

Rethinking assessment however is just one first step in the interconnectedness of higher education. The real challenge remains building an educational ecosystem that supports the shift in practice – one that includes:

• Smart and adaptive learning analytics that support providing ongoing insights rather than final judgments. This for me is the game-changer and one that I've struggled to find a sustainable and feasible solution for.

• Portfolio systems that capture development across time and contexts. UNSW eportfolio guide provides a comprehensive overview

• Faculty development frameworks that focus on adaptive mentorship. Brain Hodges' article on Learning from Dorothy Vaughan: artificial intelligence and health professionals helpfully propositions faculty to prepare themselves for the uncertainty ahead, similar to how Dorothy proactively prepared herself and her team

• Assessment designs emphasising longitudinal development and uniquely human capabilities. Adam Bridgeman Danny Liu Ruth Weeks article of a program level approach is a helpful start to thinking through this.

• Robust data governance frameworks that protect students’ privacy while enabling personalised and adaptive learning. Jason M. Lodge Tertiary Education Quality and Standards Agency (TEQSA) publication is concise and comprehensive guide

• Adaptive learning that can respond to evolving regulatory requirements across different jurisdictions while at the same time allow flexibility to embrace technology advancements. There are many platforms available but none that is fully fit for purpose. This is a wicked problem if we want to adopt a longitudinal program-wide approach. Adaptive learning platforms that could feed useful real-time data for learning analytics is a game changer. University of North Carolina at Charlotte has a helpful case study that illustrates how impactful an effective platform could be.

This systems transformation must also account for varying cultural contexts. What works in one global region may not translate directly to another. International institutions particularly must navigate different cultural expectations around assessment, varying regulatory environments, and diverse attitudes toward AI integration.

Applying a systems thinking approach to AI might transform potential threats into opportunities for educational innovation.

The true challenge isn't technological—its conceptual. Ultimately, the future of assessment isn’t about AI—it’s about redefining the student-university partnership. By embracing assessment not simply as judgement, but as a continuous, human-centered journey, universities can remain invaluable in an AI-driven world.

What strategies might help universities integrate AI as an ally in learning?What innovative assessment models are already leading the way?

Acknowledgement: Claude 3.5 Sonnet was used on the final draft of this article to provide feedback on readability...and yes, it did suggest that the article might be too long.