For most of human history, maps were made for people. They were visual, interpretative, and selective. A cartographer decided what to show and what to hide. Abstraction was not a limitation, it was the essence of the craft. But something fundamental is changing.

Today, some of the most advanced maps in the world are not designed to be read by humans at all. They are built for algorithms.

This shift, from human-oriented cartography to machine-oriented spatial infrastructure, is at the heart of a fascinating research paper by Prof. Dariusz Gotlib, a leading Polish cartographer, my scientific mentor, and a long-time family friend along with Prof. Georg Gartner (President of ICA) and Krzysztof Miksa (long-time employee of TomTom and Stellantis). Their recent publication, “HD Maps for Autonomous Vehicles: Implications for Cartographic Theory and Practice,” asks a deceptively simple question:

What happens to cartography when the primary user of the map is no longer a human, but a machine?

From Navigation Maps to Machine Maps?

Traditional navigation maps, whether paper atlases or digital road maps, were designed to support human cognition. Even Google Maps, despite its algorithmic backend, ultimately presents a visual representation optimised for human perception. 

HD (High-Definition) maps are different. They operate at centimetre/decimeter-level precision. They include:

• Exact lane geometries
• Road curvature and elevation
• Sign and signal locations
• Curbs, barriers, guardrails
• Semantic attributes relevant for driving decisions

They are not meant to be looked at. They are meant to be computed.

For autonomous vehicles, HD maps serve as a prior model of reality — allowing the vehicle to localise itself precisely, anticipate road features beyond sensor range, and make safe decisions. Sensors perceive the present. HD maps provide structured spatial memory.

And this is where the argument becomes powerful. The paper suggests that HD maps are not just a new product category. They represent a paradigm shift in cartography itself.

When Abstraction Changes Meaning

In classical cartography, abstraction was about simplification. You generalize features to make them readable at different scales. In HD maps, abstraction serves a different purpose. It is not about visual clarity. It is about machine interpretability.

Instead of symbolization for human eyes, we now design data structures optimized for algorithms. Instead of choosing colors and line weights, we define semantic layers and topological relationships.

The map becomes:

• A spatial database
• A structured model of the road environment
• A component of an AI system

Cartography shifts from visual communication to spatial systems engineering.This challenges long-standing definitions of what a map actually is.

Is a map still a map if it is never seen?

HD Maps as Infrastructure

Another important insight from the paper is that HD maps should be treated not as static products but as dynamic infrastructure.

Autonomous driving requires continuous updates:

• Roadworks
• Temporary closures
• Construction changes
• Updated traffic patterns

HD maps are therefore:

• Versioned
• Continuously maintained
• Integrated with real-time sensor feedback

In many ways, they resemble digital twins of road networks. And this is where the topic goes far beyond autonomous vehicles.

HD maps are becoming foundational layers for:

• Smart mobility ecosystems
• Cooperative perception between vehicles
• Urban traffic optimization
• Future AI-enabled infrastructure systems

The implications for geospatial professionals are enormous.

A Blind Spot in Cartography?

Despite their growing importance in industry, HD maps have received relatively limited attention within traditional cartographic theory. This gap is precisely what the paper addresses.

It argues that the academic field of cartography should not leave HD maps entirely to automotive engineers and AI developers. Instead, cartographers bring unique expertise in:

• Spatial abstraction principles
• Representation models
• Data harmonization
• Multi-scale thinking
• Conceptual frameworks of space

In other words:
Cartography still matters — perhaps more than ever. But it must evolve. The question is:

Are we witnessing the redefinition of cartography itself?

The paper suggests that we are.

Maps are no longer only communication tools.
They are operational components of intelligent systems.

For those of us working at the intersection of drones, Earth observation, AI, and geospatial technologies, this shift feels familiar. We are increasingly building spatial datasets not just for visualisation, but for automation. And that is a different mindset. We are not abandoning human-centric maps.

Instead, we are entering a hybrid era:

• Maps for people
• Maps for machines
• And increasingly, maps for both simultaneously

The future likely lies in interoperable systems where HD maps, digital twins, and human-oriented interfaces coexist.

But recognizing this transformation is the first step.

Final Thought

HD maps are more than detailed road datasets. They are a signal that cartography is entering a new phase, one where maps are no longer primarily visual artifacts but active, machine-interpretable infrastructures shaping how autonomous systems perceive and navigate the world.

For geospatial professionals, this is not a niche topic.

It is a strategic frontier.

And it is only the beginning.