top of page

A missing literacy: Why some AI demands a geographic perspective

  • May 21
  • 5 min read
To prepare students for an AI-driven future, universities must dismantle specialised silos and produce "integrators" capable of bridging physical environments with digital systems.

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 Peter Rabley, CEO of the Open Geospatial Consortium (OGC), a membership organisation dedicated to using the power of geography and technology to solve problems faced by people and the planet. As a technology executive, investor, and geographer, Peter has spent the last thirty years creating and operating geospatial businesses that map the earth to improve lives and protect the resources of our planet. 



In the autumn of 1982, Peter Rabley arrived at the University of Michigan to study geography. By his second semester, the department no longer existed. It was dismantled with no warning, no transition plan and no conversation with the students who had chosen it. This was not an isolated incident; over the next 25 years, the same decision was repeated across public universities in the US and the UK. Geography departments were merged, renamed or dropped altogether.


Beyond being a loss for the students keen to pursue the subject, the elimination of geography indicated the loss of a critical perspective and literacy: the ability to integrate physical, economic and social systems in space. This elimination has proven sadly short-sighted; in an era of artificial intelligence, this "geographic perspective" — the capacity to understand how human systems interact with physical space — is more vital than ever.


The gap between AI and world models


Today, universities are being asked to produce students and early-career professionals who can build, audit and manage AI systems. Yet — as Peter argued in his opening provocation — one central limitation of modern AI is downstream of that 40-year-old educational failure: the people building AI technology lack any grounding in spatial thinking.


The most capable AI systems can describe our world in words and pictures, but they currently cannot interpret or navigate it. Those systems cannot safely guide a vehicle through a three-dimensional physical reality or navigate emergency responders through the underground of a building without pre-existing, highly structured data. Our most advanced AI systems are currently built on text, images and code, not on a coherent understanding of physical reality.


Consequently, across the technology industry a consensus is forming around the concept of a "world model": AI that is designed to learn directly from physical, three-dimensional reality rather than just processing online data. Tech giants and research centres are identifying these models as the next major inflection point in AI, with the tech sector spending billions of dollars to acquire the foundational multi-system knowledge that used to be taught in geography departments worldwide. But building these models effectively requires talented minds shaping the inputs they are trained on.



Horizontal infrastructure as a strategic challenge 


The successor to traditional geography in the era of AI could be thought of as "geospatial architecture” — the invisible layer of location data, physical mapping and system tracking that underpins global logistics, defense, aviation, utilities and autonomous networks. Governments across the world have already committed USD $1.2 trillion to space and geospatial architectures by 2034.


But despite its global importance and massive financial backing, this remains an invisible literacy within higher education. Geospatial tools are routinely treated as a vertical budget line in IT departments or a niche elective in engineering, rather than an essential skillset that cuts across disciplines.


This oversight has occurred exactly at the moment when society urgently needs "integrators" — professionals who can bridge physical environments with digital systems. By keeping degree programs in specialised silos, universities leave most graduates ill-prepared to understand, audit or govern the technical systems shaping our world.


The importance of interdisciplinary learning


The roundtable conversation highlighted how deeply these siloed academic structures are impacting the university workforce pipeline. One engineering leader noted that while universities have successfully taught students to build "digital twins" (virtual replicas) of individual buildings for years, expanding that concept to mapping, simulating and even managing systems across large swathes of geography — even the entire world — requires a much deeper, cross-disciplinary integration.


Meanwhile, students and recent graduates feel a severe sense of urgency. As the incoming workforce, they recognise that technical skills are facing a shrinking shelf life, and they are actively demanding more experiential, social learning opportunities that teach them how to apply technical knowledge to messy, real-life situations.


Many educators are responding by focusing heavily on arts, humanities and ethics as the human counterweights to automation. For example, institutions like the Stockholm School of Economics are intentionally prioritising these "offline" human skills. This approach frames system-thinking and spatial literacy as a human-centred capability (rather than technical code to be memorised). It is precisely the cross-cutting perspective required to audit automated systems and ensure they remain governed by human experience rather than machine logic.



A 24-month window for change


The window for higher education to reclaim its agency and shape this landscape is closing rapidly. AI companies are not waiting for university curricula to catch up; they are buying, hiring or inventing their way into physical data tracking. Roundtable participants warned that without immediate academic intervention, the world models governing a rapidly increasing percentage of our physical reality will likely be entirely built, owned and licensed by just four global corporations. Concrete progress within the next 24 months is essential to better inform the people building these systems.


A powerful counter-example of how to handle this data transition comes from the province of Flevoland in the Netherlands. When developing their regional planning models, local authorities purposefully took the geospatial team out of the IT department and placed them into the policy directorate. By treating spatial systems as a question of governance and literacy rather than a software procurement problem, Flevoland successfully built comprehensive digital twins, deployed highly effective AI modelling and attracted substantial international investment.


Universities must look at their own structures through the lens of the Flevoland example, moving spatial and systems thinking out of the technical basement. We must establish "world model literacy" — the capacity to problem-solve at the intersection of physical space and digital automation — as a universal, horizontal baseline for all students.


The goal is not to resurrect historical geography departments. but to recognise that the digital data governing our physical world must be understood, structured and used in ways that are established and governed with diverse interdisciplinary perspectives. Higher education must act quickly to produce the integrators who can ensure that is the case.



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