Putting the Intelligence in Artificial Intelligence: Fireside Chat with LTG Anthony Hale on Military Intelligence Today and Tomorrow

On Tuesday, November 13th, 2024, Georgetown University had the honor of hosting Lieutenant General Anthony Hale, Deputy Chief of Staff for Intelligence (G-2) of the United States Army, on the Hilltop for the second iteration of the “Intel Speaker Series” moderated by Dr. Genevieve Lester. This speaker series focused on the growing role of Artificial Intelligence and the future role of intelligence in warfare. 

The State of Play and Military Intelligence’s Role Today

Using an acronym derived from the Army War College in the late 1980s, LTG Hale opened the fireside chat by describing today’s national security landscape as a VUCA environment—one characterized by volatility, uncertainty, complexity, and ambiguity. Due to the co-occurrence of geopolitical threats across multiple regions, LTG Hale expressed his belief that today is the most complex geopolitical environment since the dawning days of World War II. In INDOPACOM (Indo-Pacific Command) North Korea’s increased belligerence through missile testing and expansion of nuclear capabilities has created the need for continuous monitoring and evaluation by the Army Military Intelligence Corps. Regarding the PRC, Military Intelligence is faced with helping to uncover strategic intentions in the Asia-Pacific region and advancements in military technology of a highly competent adversary, in addition to more technical aspects of intelligence such as countering electronic and cyber warfare. In EUCOM (European Command) and CENTCOM (Central Command), Military Intelligence actively supports Ukrainian partners and Israeli allies with real-time intelligence reports and analysis ranging from troop movements to strategic changes in the conflict to bolster each country’s defense. 

When asked about Military Intelligence’s specific role in today’s national security environment, LTG Hale divided his answer into function, process, and product. As a function, Military Intelligence is fundamentally tasked with providing requested intelligence assessments to support land operations and warfighting. Today, Military Intelligence exists to help commanders make decisions in an ever-evolving national security landscape. As a process, intelligence is more of an art. Intelligence’s role today first requires understanding the commander’s information needs and then pulling relevant data from all sources across intelligence disciplines, including signals, human, open-source, and measurement and signatures intelligence, into a final judgment. It is then the job of the intelligence staff overseeing analysts to distill what is and is not a relevant data point to a commander’s decision-making.  Lastly, the intelligence product is a final judgment that results from collecting, processing, evaluating, analyzing, and interpreting the previously selected data to make sense of international events or relationships. A Military Intelligence professional’s judgment is paramount to how policy decision-makers arrive at an outcome, and present-day advancement in AI is one vector that enables a holistic judgment through fusing multiple sources of intelligence. 

In today’s conflicts, which span multiple regions, technology has enabled Military Intelligence to monitor developments more effectively through enhanced cyber and signals capabilities, advanced ISR (intelligence, surveillance, and reconnaissance), and open-source tools that monitor public contributions through social media and other crowdsourced platforms. Going forward, Military Intelligence must be equipped to distill these large data sets unique to multiple theaters of conflict to make informed judgments and recommendations to commanders. By incorporating AI, LTG Hale is trying to bridge the gap between these massive intelligence data points and a finished intelligence product.  

Putting the Intelligence in Artificial Intelligence 

The current hurdles for AI adoption in U.S. Army Military Intelligence stem from how data is sorted and the challenges associated with aggregating typically siloed pieces of intelligence while ensuring their security. Though automation exists to create shortcuts, the challenge for the military specifically is training the models to select the right kinds of data and eliminate bias in data selection. LTG Hale characterized the challenge of adopting AI as inherently a data problem. LTG Hale believes that selecting data, integrating and sorting data into an AI large language model, and finally, distilling the sorted data into a coherent picture, is the future of AI in intelligence but an obstacle that must be overcome. He also stated that while AI will be relied on to enhance the work of the Army’s intelligence professionals, humans must still be “on the loop.” In doing so, servicemembers must still decide what data is relevant and what data set warrants further consideration. 

To keep humans “on the loop,” LTG Hale has envisioned creating a competition amongst technology companies aiding analysts in INDPACOM which he deemed a “bake-off.” In the competition, intelligence soldiers and civilians will experiment with ways to make data selection and sorting more effective and efficient by working with technology companies to implement large language models into their daily workflow. Though LTG Hale encourages the use of technology to enhance the intelligence the Army produces, he emphasized that the intelligence professional’s human judgment must be relied upon to discern relevant information for the inherently human endeavor of war. If Artificial Intelligence is to be used in the future, it must be effective enough to pull the correct data points. After listening to the G-2, I believe it’s feasible that AI has the potential to become the ultimate all-source “-INT”––the intelligence discipline that draws from all forms of intelligence to come to a central judgment. Going forward, AI can excel by allowing intelligence service members and civilians to become data executioners, rather than data collectors, who decide the fate and importance of relevant points when evaluating all-source information. 

Though AI can facilitate decision-making, it can never be the decision-maker. While AI must make algorithmic judgments to categorize information effectively, an ongoing discussion in the national security community ensures AI is not the lethal arbiter. During the section that focused on ethics in AI, LTG Hale again underscored the importance of humans as the principal actors in intelligence. While AI can enhance intelligence produced by intelligence professionals, they must make sense of the data, and commanders must ultimately utilize intelligence recommendations and products to engage and destroy the enemy. LTG Hale went on to advocate for governance in AI without hampering innovation. This seemingly fine line can be walked by installing guardrails for intelligence professionals and the private industry to innovate within, thus adhering to the law but still having room to drive innovation forward. The guardrails the G-2 was referencing would likely ensure human oversight for critical decision-making involving life and death rather than the use of automation. Additionally, regulatory frameworks enable auditing and monitoring of AI as well as access controls ensuring only authorized users can experiment with AI for military applications. These boundaries would allow the military to innovate without fear of third-order consequences.

Ultimately, LTG Hale forecasts that AI—if used correctly—will change how the consumer handles intelligence. An underlying theme of the fireside chat was the need for intelligence professionals to have command over the inputs and outputs AI is producing. While AI can help sort data, intelligence professionals must still be the executors determining what is relevant. To address AI’s potential opportunities and shortfalls in intelligence, LTG Hale emphasized increased data literacy amongst those he oversees and the timeless need for human judgment in intelligence analysis. Though AI will enhance the work they do, intelligence professionals must exist to think critically about intelligence issues leveraging history, personal experiences, and training as a foundation.

Views expressed are the author’s own and do not represent the views of GSSR, Georgetown University, or any other entity. Version approved for public release. Image Credit: INSA Online