University of Illinois Chicago

Szymon Machajewski is Associate Director of Academic Technology and Learning Innovation at the University of Illinois Chicago, while Mike Harding leads the Office of Digital Accessibility at California State University, Fresno. Together, they bring expertise in digital learning, instructional innovation and digital accessibility across higher education.

When GPS replaced the paper map, something changed

A generation grew up without learning how to unfold/fold a map, visualize the relationships between neighboring towns, or navigate by landmarks and cardinal directions. Yet GPS delivers clear benefits: fewer wrong turns, real-time traffic information and safer travel. The tradeoff seemed reasonable because the gains were immediate and tangible, while the losses were subtle. AI presents a similar bargain. This time, the territory is not geography. It is knowledge itself.

In higher education, discussions about AI often return to the same questions and recurring debates. A common response has been prohibition: restricting emerging technologies and requiring students and faculty to adhere to established practices. As the authors of this article, working across academic technology, digital accessibility and teaching and learning, we contend that this response misses the central issue. Prohibition preserves existing processes, but risks obscuring the broader educational mission.

Learning is inherently cumulative - it builds on itself through engagement with and extension of what has come before. From this perspective, the real question is not whether higher education should use AI. Rather, it is whether AI can be leveraged to expand our capacity for thinking, learning and intellectual growth, or whether its use gradually diminishes the cognitive resources on which those capacities depend.

We refer to this challenge as cognitive capital sustainability. Borrowing from environmental sustainability, the concept recognizes that valuable resources can be cultivated or depleted over time. Cognitive capital consists of the habits of reasoning, knowledge, judgment and intellectual skills accumulated across generations. Unlike material resources, cognitive capital exists only through active use.

Physical training provides a useful analogy. Like muscles, cognitive abilities strengthen through challenge and weaken through disuse. Thinking is not a fixed capacity to be preserved but a capability that develops through effort. AI used primarily to bypass intellectual work represents a form of cognitive extraction. Consider a robot that lifts weights on a person’s behalf at the gym: the movement occurs, but the intended benefit is lost. The same principle applies when AI is used as a substitute for the cognitive work required to evaluate, question and improve.

Used differently, however, AI can function as a training partner rather than a substitute. It can help us challenge assumptions, expose potential weaknesses in an argument, introduce alternative perspectives and push us towards problems that are more complex. The critical question is whether AI is reducing cognitive demands or helping us develops the capabilities needed to meet them.

Answering that question requires applying a simple test - distinguishing between tasks that warrant human effort and those that do not. When AI is used to automate routine tasks such as formatting, scheduling, or filing, the cognitive cost is minimal. When AI replaces judgment, interpretation, creativity, or original argumentation, the costs are far, far greater. These remain essential human capabilities and are precisely the forms of cognitive work we should seek to cultivate.

The same principle appears in digital accessibility: Strip away the barriers that interfere with equality and learning, preserve the intellectual challenges that constitute learning itself. Automate what is incidental so that greater attention can be devoted to what is most essential.

There is a blind spot here that higher education leaders overlook. "AI strategy" is frequently treated as one decision when it is really, in fact, several. Teaching responsible AI use, improving operational efficiency, redesigning coursework and preparing students for professions that are currently being transformed by AI are related but distinct challenges. Most institutions will focus on the first two because Addressing these challenges does not require higher education professionals to become experts in every new AI system. If anything, surviving and evolving in higher education and the professional horizon ahead requires a renewed commitment to intellectual adaptability. Productive response to technological change is not mastery of every emergent tool but cultivating the willingness to learn publicly, revise our assumptions and model curiosity. An expert who cannot become a beginner again is an expert with an expiration date.

The cognitive capital sustainability challenge of this nascent AI era is not technological. It is human. GPS did not make us worse at arriving. It replaced one way of getting there with a faster, safer one and gave us back the attention we used to spend on the map. Nobody misses being lost. Can the cognitive effort that AI removes from work be reinvested in higher-order thinking and creativity? Perhaps the future value of AI in higher education depends less on what the technology can do than what the educators, staff and students can do with the cognitive capacity it creates.