OGC standards are used by governments, businesses, and academic institutions worldwide to create and manage geospatial data, and build enterprise applications and services, that work behind the scenes to make the modern world work!

Standards development organisations (SDOs) have been essential in bringing order and interoperability to complex domains. They provide a framework for creating and maintaining standards that ensure different products and services can work together seamlessly. At one level this is interoperability but there is more to it…

Convening Power: Bringing the Experts Together

One of the most important values of SDOs is their ability to convene and engage a diverse community of stakeholders. The OGC, for instance, brings together representatives from a wide range of organizations, including software developers, data providers, government agencies, and research institutions. This diverse participation ensures that a broad spectrum of perspectives and needs are considered in the standards development process.

The OGC’s convening power is particularly important because it brings together experts from different backgrounds and with different interests. This includes both competing commercial companies and government and academic users. This diversity of perspectives helps to ensure that the standards produced are not only technically sound, but also practical and meet the needs of a wide range of users.

Expanding the Circle: The New Individual Membership

To further enhance this convening power, the OGC recently announced a groundbreaking new class of membership: the Individual Membership. Officially launched at the Philadelphia Member Meeting, this initiative opens the door wider, providing a direct pathway for independent developers, consultants, professionals, and students to participate directly in working groups, code sprints, and testbeds.

This new membership class brings the potential for a significantly more diverse community. Interoperability improves when more implementers can share what they are seeing and surface edge cases. By lowering the barrier to entry, the Individual Membership aims to amplify voices from historically underrepresented regions—such as Latin America, Africa, and Southeast Asia. Broader geographic and professional participation makes the work more resilient, ensuring that an even wider range of perspectives helps shape the future of geospatial standards.

Process: Creating Confidence in the Standards

Another key value of the OGC is the rigorous process they follow for creating and maintaining standards. The process is designed to ensure that the standards are high quality, relevant, and supported by a consensus of the community. The title of this post “Any Objection to Unanimous Consent” comes from the Robert’s Rules of Order and is a commonly heard phase during meetings, representing a checkpoint for consensus – key to the process.

This process typically involves multiple stages, including:

1. Working Groups: OGC working groups are formed to develop and maintain standards for specific areas of geospatial technology. These working groups are composed of experts from a wide range of organizations.
2. Public Review: Once a draft standard has been developed, it is released for public review. This allows the broader community to provide feedback on the standard and ensure that it meets their needs.
3. Approval: After public review, the draft standard is submitted to the OGC membership for approval. A majority vote is required to adopt the standard.
4. Maintenance: The OGC provides ongoing maintenance for its standards. This involves updating the standards to reflect changes in technology and addressing any issues that may arise.

This rigorous process provides confidence in the standards and reports produced by the OGC. It ensures that the standards are technically sound, meet the needs of the community, and are supported by a consensus of the relevant stakeholders.

Part of my role and the role of the board and the leadership team of the OGC is to make sure that this process remains fit for purpose and relevant in the age of vibe coding and AI – Watch this space !

The OGC is a prime example of an SDO that creates significant value for its constituents. Its convening power, expanding community diversity, and rigorous processes ensure that the standards produced are high quality, relevant, and supported by a consensus of the community. The work of the OGC remains essential for enabling the effective use of geospatial information and for building a more connected and interoperable world.

In an era where our smartphones are essentially extensions of our physical selves, the question of privacy—specifically location privacy—has moved from a niche technical concern to a primary user priority. The psychology of digital trust is key to bridging the gap between providing useful, localised services and respecting the inherent sensitivity of a user’s physical movements – A Digital Fingerprint of unique importance.

The Sensitivity of “Where”

Unlike a username or an email address, location data reveals a person’s life in real-time: their home, their workplace, their hobbies, and even their health habits. Because this data is so personal, the “trust threshold” for location-based services is significantly higher than for other types of digital interaction.

Smartphone users are no longer blindly clicking “Allow.” Instead, they are performing a rapid, often subconscious cost-benefit analysis. They ask themselves: Is the value of this localised content worth the potential risk to my privacy? 

To win this internal debate, location platforms and app developers must focus on both transparency and perhaps most vitally immediate relevance.

Signals of Trust

How does a user decide an app is trustworthy? Users are becoming sophisticated in their understanding of locational privacy, perhaps more than we in the industry appreciate. I would argue this understanding is the result of the best practice the industry has largely adopted over the last decade or so.

There are several key “trust signals” that influence user behaviour:

• Visual Cues and Timing: Trust isn’t built through the fine print of an app’s “Terms of Service”; it’s built through interface design. Clear icons, brief notices, and confirmation messages that appear at the moment the data is needed—rather than in a “blanket request” at first launch—help reassure users.
• Plain Language: Technical jargon and “legalese” are trust-killers. Users respond far more positively to short, plain-language explanations that answer the simple question: “Why do you need to know where I am?”
• Neutral Tone: Persuasive or pushy language (“You must enable location for the best experience!”) can trigger a defensive response. In contrast, neutral, explanatory language feels more respectful of the user’s autonomy.

Permission Design

One of the most effective ways to build confidence is by offering granular control. When a user is given the choice to “Allow Once” or “Only While Using the App,” they feel empowered rather than cornered. This shift from all-or-nothing data collection to situational access is a cornerstone of modern privacy standards (like those seen in recent Android and iOS updates) and is essential for maintaining long-term user engagement.

Where the industry could do better is in consistency. If a system repeatedly asks for the same permissions after being denied, or if settings seem to change without user input, trust evaporates. Reliability across sessions creates a “safety net” that encourages users to return.

Transparency and Context Relevance

One of the recurring points in the discussion on Trust in GeoAI was Transparency.

Transparency should not be a one-time event during the onboarding process. Privacy information should be easily accessible at all times, allowing users to review how their data is being used whenever they feel the need. This ongoing transparency transforms a transaction (data for service) into a relationship.

However, even the most transparent app will lose a user’s trust if the content delivered isn’t relevant. There needs to be a balance established between utility and restraint. Over-targeting—showing too much location-specific detail too quickly—can feel “creepy” and invasive. For instance, if an app knows exactly which aisle of a store you are in before you’ve even expressed interest in a product, the relevance is overshadowed by the intrusiveness. The goal is to provide content that aligns with the user’s intent without crossing the line into surveillance.

The marketplaces of both iPhone and Android ecosystems still contain a large selection of Flashlight apps that request the user’s location. There is, of course, no justification for this other than to support local advertising and data gathering – most users recognize this but sometimes “break the glass” because they are somewhere dark without a torch. There is no trust, but sometimes app usage is purely transactional?

Autonomy: Opting In and Out

Trust is deeply linked to the ability to walk away. Systems that make it easy to opt-out of location tracking, or that provide a “private mode” for exploring sensitive topics, demonstrate a respect for the user that pays dividends in brand loyalty.

Whether a user is looking for a local restaurant or navigating more sensitive local listings, they want to know they are in the driver’s seat. Systems that prioritise discretion and allow users to explore options without being forced into a data-sharing agreement are ultimately the ones that will thrive in a privacy-conscious market.

Lessons for GeoAI?

As location-based technology becomes more sophisticated, the “creepy factor” remains the biggest hurdle for developers. Trust is not a static checkbox but a conversation, lasting as long as a user makes use of an app or service. By combining thoughtful permission design, plain-language transparency, and respect for user autonomy, platforms can provide the localized experiences users want without making them feel like they’ve sacrificed their privacy to get them.

Despite location technology having the characteristics of a magic and invisible force, and for most users a system that’s operation is not fully understood, society does now have an understanding and expectation of “how” the system operates both technically and, as importantly, in terms of business models – location sharing is a two-way street!

The same cannot be said for many AI systems based on foundation models; by their nature, they remain “black box” systems. While deriving their capability from training data, how that data is used to respond to a user interaction or prompt is non-deterministic.

Transparency is key in building trust for location technology but may not be enough for GeoAI; explainable GeoAI still needs some work.

Update – 05/02/26

This post from the MIT Technology Review is interesting while focused on the security of Agentic AI systems, the approach of treating AI systems as Human Actors is also appropriate to viewing issues of Trust. Would you trust the intern to write a client report without supervision ?

Happy New Year!

Remember “WebGIS”? The term itself feels a little anachronistic now, doesn’t it? It conjures images of a distinct, almost revolutionary movement, a time when putting maps and spatial analysis on the internet was a novel and exciting frontier.

I know I was there…

Today, the very idea of differentiating “WebGIS” from “GIS” is almost quaint. Geospatial technology is fundamentally web-enabled. The internet isn’t an add-on; it’s the foundational architecture upon which modern GIS operates. This natural, pervasive integration offers a powerful parallel for understanding the trajectory of another transformative force: Artificial Intelligence.

For the past few years, we’ve seen the rise of “GeoAI” – a term that describes the intersection of AI and geospatial technology and one which you all know I find troubling (why do we will compelled to put “Geo” in front of anything new??) . It encompasses everything from machine learning for feature extraction from satellite imagery to deep learning models for predicting urban growth or optimising logistics.

And much like “WebGIS” before it, “GeoAI” does however mark a significant paradigm shift. It highlights the new capabilities, the new research directions, and the new challenges that emerge when AI is brought to bear on spatial data.

But just as “WebGIS” eventually faded into the background as web integration became the default, so too will “GeoAI” likely become a historical term. This isn’t to diminish its importance or the incredible innovations it represents. Quite the opposite. The eventual disappearance of “GeoAI” from our lexicon will be the ultimate testament to its value.

Think about it: when we talk about GIS now, we inherently assume a web-based infrastructure. We expect interactive maps, cloud-hosted data, and collaborative tools that leverage the power of the internet. We don’t append “web” because it’s no longer a distinguishing feature; it’s simply how things are done. Yes there are of course a few edge cases where workstations are still used but the majority of us are sitting in front of a browser most of the day.

The same destiny awaits AI in the geospatial realm. As AI algorithms become more sophisticated, more accessible, and more deeply embedded into every facet of geospatial workflows, the need to call it “GeoAI” will diminish.

We won’t be talking about “AI-powered mapping” as a special category; we’ll simply be talking about “mapping,” with the understanding that intelligent automation and analytical capabilities are an intrinsic part of the process.

Imagine..

Imagine a future (well it’s not really the future ) where a satellite automatically identifies and classifies objects as it orbits, sending back not imagery but processed, intelligent data that is immediately actionable. 

Where a city planning model dynamically optimizes infrastructure based on real-time traffic, demographic shifts, and environmental factors, all powered by unseen AI algorithms.

Where your everyday mapping application provides predictive routing based on learned patterns of congestion and even suggests new points of interest based on your preferences and historical movements.

These are not just “AI features”; they are the core functions of a truly intelligent geospatial system.

A journey well traveled..

The journey from “WebGIS” to just “GIS” taught us that foundational technological shifts eventually become so pervasive that they shed their distinguishing labels.

They simply become the new normal. “GeoAI” is currently doing the vital work of pushing boundaries, inspiring innovation, and attracting talent to this exciting interdisciplinary space. But its greatest legacy will be its own obsolescence – a sign that intelligence has not just intersected with geospatial technology, but has become its very fabric.

The future of Geospatial isn’t “GeoAI”; it’s simply Geospatial, made infinitely more powerful and insightful by the intelligence woven into its core.

Now if only we could get rid of the term Geospatial and go back to using Geography !

A Guest Blog by Larry Vaughn, Former Mayor of Amity Island

You know me. I’m the guy who wore the anchor-print suit. I’m the guy who told the press, “I’m pleased and happy to repeat the news that we have caught… and killed… a large predator.” I’m the guy who kept the beaches open on the Fourth of July because I was terrified of losing those summer dollars.

And yes, I’m the guy who was wrong.

I learned a hard lesson in 1975: ignoring the experts because the reality is inconvenient doesn’t make the problem go away. It just makes the consequences messier.

Looking at the modern business landscape, particularly in the Geospatial Industry (if it is one?) , I see a lot of you making the same mistakes I did. You have more data than that pesky Chief Brody ever had. You have satellites, drones, and location intelligence that can track a crab across the sand from orbit. But when it comes time to make the hard call, you’re still standing on the ferry acting like everything is fine.

Here is why leaders are drowning in data but starving for wisdom, and why your maps are being ignored just like I ignored that chewed-up girl on the beach.

The “Summer Dollars” Syndrome

In Amity, the logic was simple: If we close the beaches, the town dies. The economic data (short-term profit) outweighed the biological data (there is a giant shark).

In the geospatial world, you collect massive amounts of information. You have terabytes of raster imagery, point clouds, and vector data. But often, leaders overlook this data when it conflicts with their “gut feeling” or immediate quarterly goals.

Take Town Planning and Flood Risk. We have LIDAR data now that can map elevation changes down to the centimeter. We can model exactly where the water will go during a 100-year storm. The GIS analysts (your modern-day Matt Hoopers) point to the map and say, “Mr. Mayor, if you build those luxury flats here, the basement will be underwater in five years.”

But the developer sees the waterfront view. The Council sees the tax revenue. So, what do they do? They say, “You’re gonna yell ‘Barracuda!’ and panic everyone?” They ignore the hydrological model, approve the development, and five years later, they’re asking for a government bailout. They prioritised the “summer dollars” over the geographical reality.

Treating Data Like a “Bad Fish”

Remember when some guys caught a Tiger Shark and we all thought the nightmare was over? Hopper told me, “The bite radius is different.” He had the metrics. He had the forensic measurement. I didn’t want to hear it because the Tiger Shark was an easy answer.

Leaders do this with Site Selection all the time.

I’ve seen retail giants collect petabytes of mobile location data. They have heatmaps showing exactly where their customers live, drive, and shop. The data says, “Open the new distribution center in Sector A to optimize delivery times by 15%.”

But the CEO? He likes Sector B better. Maybe it’s closer to his golf course. Maybe he just has a “good feeling” about it. He looks at the heatmap and treats it like a suggestion rather than a science. He ignores the spatial correlation because it doesn’t fit the narrative he wants. He hangs the Tiger Shark on the dock and calls it a victory, while the Great White is still swimming in the P&L statement.

The Glitch in the Human Operating System

Here is the thing nobody admits: Maybe I wasn’t just being greedy. Maybe I was suffering from a Man-Machine Interface failure.

We humans—even Mayors—are irrational by nature. We are wired to seek patterns that confirm our hopes (optimism bias) and ignore patterns that confirm our fears (normalcy bias).

In your industry, you build incredible dashboards. You create Digital Twins of entire cities. But have you considered the “user interface” of the decision-maker’s brain?

When you hand a non-technical leader a complex GIS map layered with fifty variables, their brain often shuts down. They can’t process the signal from the noise. It’s too abstract. So, they revert to their default setting: Irrational Hope.Hooper showed me science; I saw a complication. The “machine” (the data) was working perfectly, but the “operator” (me) couldn’t parse the output.

If your geospatial insights aren’t translated into a language that cuts through human irrationality—if it’s just raw data without a compelling narrative—your leader is going to stare at it, blink twice, and say, “I think the beaches are safe.”

Listen to Your Chief Brody

Your GIS team, your data scientists, your remote sensing experts—they are the ones out on the boat chumming the water. They are the ones seeing the blips on the sonar.

When they bring you a dashboard showing that your supply chain is vulnerable to climate risks, or that your agricultural yield prediction is down based on multispectral satellite imagery, don’t gloss over it.

I know it’s tempting. I know you want to shout, “Amity is a summer town! We need summer dollars!” I know you want to launch the product anyway, or build the factory on the fault line, or ignore the demographic shift because it’s inconvenient to pivot.But spatial data is the most grounded reality you have. It literally maps where things are happening and why.

Conclusion

If I could go back to ’75, I would have signed that order to close the beaches. I would have listened to the guy who knew sharks, not the guys who knew tourism.

Don’t be the Larry Vaughn of your industry. You spent millions collecting geospatial data. You have the map. You have the coordinates. Don’t wait until the shark comes up and bites a hole in the bottom of your boat to start paying attention to it.

Stay safe, and check your maps.

Larry Vaughn is a fictional character from the 1975 Jaws motion picture. Merry Christmas Everyone !

Enter the hot term of the moment,  geospatial embeddings, a core innovation of modern Artificial Intelligence. These high-dimensional vectors are now becoming a foundation of the field, promising to unlock insights at planetary scale.

However, while we embrace this powerful new capability, we must also critically examine the sacrifice of transparency and trust in favour of efficiency.

What Are Geospatial Embeddings, and What Do We Lose?

At its most fundamental, an embedding is a data compression and transformation technique. It converts highly complex, multi-temporal EO data— both the spectral and contextual history of a single spot on Earth—into a compact, continuous, numerical vector.

This is achieved using massive Foundation Models trained on trillions of pixels globally. These models it’s claimed extract the “semantic” meaning of the data, not just what a location looks like right now, but what it is (e.g., ‘a coastal mangrove forest’) and how it behaves (e.g., ‘undergoing gradual drought stress’).

The claims are indeed impressive: similarity of landscapes in the real world equals proximity in the vector space.

However, the sacrifice is the direct physical interpretability that traditional remote sensing was founded on. A simple Near-Infrared value tells you about leaf density, while an embedding dimension tells you… nothing on its own.

We supposedly gain “semantic” meaning, but we trade it for algorithmic opacity.

I appreciate this might me difficult to follow, perhaps a little nostalgia to my formative years in Remote Sensing might help..

Back to the future.. Image processing 1980’s style !

When I was a boy, or at least a Masters Student of Applied Remote Sensing, when it came to segmenting an image or classification to use the term of the day, you had the choice of using special signatures at various wavelengths to train a classification algorithm – supervised classification or using algorithms to identify similar groups of pixels automatically – a technique know as unsupervised classification.

So the concept of grouping similar features isn’t new. For decades, analysts relied on unsupervised classification using algorithms like K-Means. However I would argue although less transparent than supervised classification, unsupervised classification was still superior to embeddings in terms of explainability.

A comparison reveals how fundamentally the connection between the analyst and the data has changed.

The move to embeddings offers a necessary efficiency gain, but we should acknowledge that we are embracing a system that outsources the core act of feature engineering to a black box. While unsupervised classification was cumbersome, its output (a handful of spectral clusters) was fully auditable and understandable by a human analyst.

Embeddings are powerful precisely because they are cryptic.

Spatial Autocorrelation Ignoring the law of Geography ?

Perhaps the greatest internal challenge facing the adoption of geospatial embeddings lies in the very nature of geographic data itself: spatial autocorrelation.

Spatial Autocorrelation  is the formal expression of Waldo Tobler’s First Law of Geography: “Everything is related to everything else, but near things are more related than distant things.”

In the context of Earth Observation, this means that a pixel representing a healthy forest has an attribute (its spectral signature and, consequently, its embedding vector) that is highly similar to the pixel immediately next to it. Features on Earth—like a large field, a mountain, or a lake—don’t stop abruptly at a single pixel; they extend continuously across space.

This of course sounds logical, but for AI models this is both a blessing and a curse..

The Blessing (Efficiency): Deep learning models, particularly Convolutional Neural Networks (CNNs) used to generate embeddings, thrive on spatial autocorrelation. Their convolutional filters are designed to look at a neighbourhood of pixels to determine the context of a central pixel, effectively enforcing Tobler’s law by ensuring that the resulting embedding vector captures local spatial dependencies. This is what makes the embeddings so rich in context.

The Curse (Statistical Integrity) : Spatial Autocorrelation poses a significant challenge to statistical integrity in geospatial data. In classical statistics and machine learning, we assume data points are independent and identically distributed. However, this assumption is often violated in geospatial data particularly when we combine spectral EO data with other contextual data such as elevation or temparture data. This leads to two critical issues in embedding-based workflows:

1. Inflated Model Performance: When models are trained or validated using samples too close together due to spatial autocorrelation, the high similarity results in overly optimistic performance metrics. The model doesn’t learn a generalisable rule but rather memorises local spatial trends, leading to the false belief that an embedding is highly predictive when applied to distant, dissimilar regions (spatial heterogeneity)
2. Sampling Bias: Analysts must carefully de-correlate their samples when creating ground truth labels for downstream models (e.g., classifying land cover based on pre-computed embeddings). If samples are taken too close together, they contain redundant information, skewing the training process and failing to capture the true global variability of the phenomenon.
   

Show your workings..

Remember exams when you were asked to explain out you got to you answer…

An embedding is a point in a vast, high-dimensional space: 

If an AI model, using this vector, flags a specific region as “High Risk for Illegal Deforestation,” policy makers and conservationists need to know why. What specific real-world feature does the value of -0.81 in the 32nd dimension correspond to?

The blunt answer is: it doesn’t correspond to any single thing. The intelligence is collective.

This has all the makings for a crisis of trust in politically sensitive applications of geospatial information in particular at a global scale..

• The Scientific Audit: Unlike traditional spectral indices, embeddings cannot be easily validated using ground truth data due to their nonlinear and opaque input-output relationship. This makes it challenging to verify the scientific rigour of insights based on untraceable fundamental feature representations.
• Bias and Fairness: If the foundational model was primarily trained on data from the Global North, its embeddings might poorly or unfairly represent landscapes, agricultural practices, or built environments in the Global South. This algorithmic bias is difficult to detect and correct within the black box of the vector space.
• Regulatory Compliance: As governments increasingly rely on EO-derived intelligence for policy and enforcement, a critical question arises: can an AI prediction based on an uninterpretable embedding vector withstand legal scrutiny? The requirement for a human-understandable chain of evidence remains a significant roadblock.
  
  

Conclusion

The path to reliable, globally robust geospatial AI lies not just in bigger models, but in models that respect, measure, and account for the fundamental geographic laws that govern our planet.