GEOINT, AI and Analysis: Achieving Speed to Mission with Human Analysts as Guardians
Three Points to Remember
- AI-enabled GEOINT offers powerful advantages for agencies across the U.S. government.
- Intelligence analysts act as “guardians” when using AI to ensure insight and context in final products.
- Analysts must hone their skills in three key areas: analyst tradecraft, debate and expertise.
Today’s threat landscape moves fast and rarely looks straightforward. Adversaries act quickly, flood systems with data, and operate across multiple domains to hide their intentions and force faster decisions.
In response, geospatial intelligence (GEOINT) tradecraft has moved to the forefront of AI adoption. AI-enabled GEOINT offers powerful advantages for agencies across the U.S. government, but only when paired with intentional human analyst oversight to ensure trust and mission relevance. The National Geospatial-Intelligence Agency (NGA) has been leading this shift for years.
As a 25-year veteran of the U.S. Intelligence Community and an analyst by training, I personally have observed the Intelligence Community taking a far more aggressive tack toward AI adoption than in the past, and indeed, perhaps more than most IC partners. This makes sense given the complexity of processing, analyzing, and disseminating the increasing terabytes of geospatial data, and for NGA, perhaps the most data of any intelligence agency.
Analysts are at the forefront of adoption
While officers across the U.S. government are ramping up their use of AI, perhaps nobody stands to gain from—and experience more interaction with AI—than analysts. Having worked alongside GEOINT analysts as a peer, supervisor and customer since the beginning of this century, I have experienced the excitement and angst that technological advancements have brought to the fore, but nothing has caused as much strife as the introduction of AI.
Perhaps the most profound and notable comments I hear from analysts today is that they want AI to readily provide three of the “Four Ws”—the ”what”, “where” and ”when”—so that they can spend their time on the ”why”—and ultimately the “so what” for the customer.
While it’s ideal to log onto a workstation and access a dashboard already loaded with critical images from AI software overnight, there is the risk of AI producing a critical error. In response, NGA announced that it was the first of the 18 intelligence agencies to create products that carry a warning label of “machine-generated”—indicating that the items have not been touched by human hands. At the time NGA Director Vice Admiral Frank Whitworth noted, “the AI itself needs human help… not only to double-check its final output, but to help train it for what to look for in the first place.”
Equipping “analyst guardians” for GEOINT AI success
We can reach an ideal state where AI generates the “what”, “where,” and “when,” while human analysts focus on the “why” and the “so what.” This balance is achieved by intentionally designing AI to support, not replace, human judgment. In this model, analysts act as guardians of insight, applying context, experience and critical thinking to extract the “last mile” of AI’s value.
Going one step further, we must empower analysts with tools, techniques and training that ensure the resulting data is trustworthy, insightful and offers U.S. customers an unfair advantage. To do so, industry can support agencies, like NGA, in honing their skills in three key areas: analyst tradecraft, debate and expertise.
- Analysis tradecraft has never been so important in ensuring precise, accurate and compelling insights when AI is generating the baseline. One of the core principles of good intelligence analysis is the coordination process that includes the sharing of draft finished intelligence products with colleagues and peers for comments and critique prior to edit and publication. One of my formative moments as an analyst was during a major international crisis while producing the President’s Daily Briefs, or PDB, that required IC-wide coordination; amid a moment of despair with the process, an astute manager rightly noted that coordinating a product rarely made it worse and almost always made it better. Effective coordination is the most basic tradecraft example that can arm the “analyst guardian,” combined with structured analytic techniques laser focused on precision of language, and an unyielding commitment to the rigor of substantiation.
- Debate is at the center of that coordination process and plays an even deeper role in ensuring analytic rigor and customer impact, and this too should be an arrow in the “analyst-guardian” quiver. No matter how pristine the product, debate around the story and its importance is key.
- Expertise is a cornerstone of intelligence analysis and will be a foundation to the increased focus on the” so what” as AI tackles the front end of intelligence questions. The definition of expertise in GEOINT has evolved from cartography, mapping and photo interpretation, to an interdisciplinary field that includes data science, human geography, AI and increasingly, the critical thinking to make sense of the full picture. Ultimately, it takes an engaged expert to add essential context to intelligence.
