From tracking climate change resilience in African drylands to monitoring crop health in vineyards, geographic artificial intelligence — or GeoAI — is fundamentally transforming how we understand our planet.
In a new book, Remote Sensing, Big Data, and GeoAI, Erin Bunting, an expert in remote sensing and leveraging big data, geospatial analysis, modeling, and statistics to understand environmental change, climate impacts, and human–environment connections from the Michigan State University Department of Geography, Environment and Spatial Sciences, explores the power, ethics, and future of this emerging field.
Below, she explains how satellites, big data and AI are moving geography beyond static maps and why critical thinking is more important than ever.
For those outside the field, how would you describe the moment we are in with satellite data and AI?
For decades, remote sensing relied on taking a single satellite image, classifying the land cover/use and describing basic patterns of change. But we were always limited. We were left asking: Is this land-use change significant? What drove it? How resilient is a specific species when our classifications only distinguish "forest" from "wetland"?
The tipping point came when satellite imagery repositories reached decades' worth of data. Suddenly, we had dense time series from platforms like Landsat and MODIS, but traditional statistical methods weren't equipped to process it all.
AI and machine learning are the bridge. They give us a suite of tools to extract meaningful information from decades of data, breaking down the subtle patterns and drivers of landscape change in ways that were previously impossible.
AI feels like it’s everywhere right now. What makes "GeoAI" fundamentally different from tools like ChatGPT?
Spatial is special! GeoAI is built specifically to understand where things happen, how places are connected, and why location matters.
ChatGPT can tell you what happened, but GeoAI helps answer where it happened, why it happened there, and what might happen next. It combines AI with spatial thinking, analyzing satellite imagery, drone footage, GPS data and climate patterns to uncover complex relationships.
It’s like looking at a city from an airplane. You can see roads, neighborhoods, rivers, parks, and buildings, but your brain naturally starts asking why certain neighborhoods are hotter or why floods repeat in specific spots. GeoAI recognizes those complex, multi-layered relationships across landscapes and over time.
How is GeoAI helping us "see" environmental changes that the human eye might miss?
Our eyes are great at spotting obvious shifts, but they struggle with subtle changes spread across thousands of square miles over many years. That's where GeoAI shines.
It acts like a super-powered set of eyes, detecting declining forest health before visual damage appears or tracking tiny shifts in wetlands over time. In my own research, we’ve used GeoAI to monitor vineyards, picking up early changes in plant health so growers can make faster, better-informed management decisions. It doesn't replace human expertise — it scales it up.
Your research focuses on climate change resilience in Southern Africa. How does big data change how we study these complex ecosystems?
In dryland ecosystems like the Okavango, Kwando and Zambezi catchments, you have a delicate balance of changing climate patterns, unique vegetation, wildlife and human communities.
Using GeoAI methods like Random Forest and Dynamic Factor Analysis, we can isolate how variables like temperature and soil moisture interact across the . With this amount of spatial data, we can pinpoint critical climate thresholds — the exact tipping points where an ecosystem shifts — and track how resilient certain land covers are over time.
We often hear that we are "drowning" in satellite data. What is the biggest challenge in turning this data into useful insights?
The biggest challenge isn't the volume of data — it's making sure we use it correctly. Today, anyone can download satellite imagery and apply AI tools in a few clicks. That’s exciting, but satellite data aren't just pictures; they are scientific measurements that require careful calibration.
Without a strong foundation in remote sensing, it’s easy to draw inaccurate or misleading conclusions. AI can analyze data at incredible speed, but it can't replace an understanding of how the data was collected or whether the results make scientific sense.
As AI becomes more autonomous in analyzing geographic data, what is the top ethical concern that keeps you up at night?
People usually think of privacy first, which is a major concern. But what keeps me up at night is false confidence.
AI is becoming so good at producing polished answers that people may stop questioning if those answers are actually correct. An AI-generated map can look authoritative even if it’s built on biased data, flawed processing, or incorrect assumptions. If decision-makers trust those outputs blindly, it impacts everything from disaster response to public health and urban planning. AI should enhance our decision-making, not replace critical thinking.
How are you updating the geography curriculum at MSU to ensure students are critical thinkers rather than just "button-pushers"?
Over the last three years, Beth Weisenborn, director of our onGEO online program, and I have completely rebuilt our geospatial curriculum at MSU with one primary goal: creating spatial problem-solvers.
We don't use simple step-by-step lab manuals where students just click through menus. We intentionally design labs around "discrepant events" — situations where the expected answer doesn't immediately make sense. This forces students to investigate, ask questions and problem-solve. Software tools will come and go, but the ability to think spatially and analyze data critically will never go out of style.
What is your advice for students entering environmental science today — do they need to learn to code?
My answer might be slightly controversial, but yes: students should become proficient in at least one programming language, like Python or R.
Is coding strictly mandatory for every single role? No. But does it make you more adaptable, competitive, and employable? Absolutely. You don’t need to become a computer scientist, but knowing how to work with data and automate tasks gives you a massive advantage. Combine strong geographic knowledge with computational tools, stay curious and always focus on the why behind technology use.