The development of sensors, rapid advances in algorithm models, and significant improvements in computing have expanded the opportunities for AI and its use in the defense industry. As a result, NATO has a unique opportunity to revolutionize its military and strategic activities.
The alliance has identified seven key areas for AI development: recognition, conversation, and human interaction, predictive analysis and decisions, goal-driven systems, patterns and anomalies, hyper-personalization, and autonomous systems. Autonomous systems act as a unifying thread between the categories, streamlining decision-making and operations in each, with varying degrees of human interaction.
There are three degrees of autonomy, commonly referred to as:
- in-the-loop, where humans have a great degree of control
- on-the-loop, where humans have only the ability to abort action, and
- off-the-loop, where humans have no control.
NATO members’ current systems predominantly lie in the first category, but technological advances have created opportunities to reduce reliance on human interaction.
Russia demonstrated the battlefield use of AI hyper-personalization in 2014, when special forces collected information on contacts, locations, and habits to create detailed profiles of their targets in Ukraine. In one example, the mother of a Ukrainian commander received a call saying her son was in the hospital. Believing it was legitimate, she called him and Russian forces were able to zero in on his location and kill him with an artillery strike.
While hyper-personalization is still in its infancy, it has the potential to be very effective for such decapitation strikes and reduce the effectiveness of enemy leadership. When combined with big-data analytics or self-learning, the targeting of commanders is set to increase at an alarming rate.
Another example of battlefield implementation of AI is the use of short-range ballistic missiles able to strike targets hundreds or even thousands of kilometers away. At a cost of roughly $1m, they are often used against fixed targets of either high economic or military value. With the addition of an AI-controlled maneuverable re-entry vehicle, they can course-correct to avoid anti-missile defenses or strike moving targets. Their most notable use is against warships at sea, greatly increasing the potential of naval anti-access/area denial (A2/AD) strategies.
Integrating AI and autonomy into NATO requires a united effort to standardize and regulate on-the-loop technologies, while harnessing the public and private sectors to enhance innovation and education. A NATO-sponsored competition, for example, might encourage AI development, drawing on the Ansari and Watson XPrizes as inspiration.
The alliance could use tech competitions to spur and source emerging technologies, with jury panels bringing together venture capitalists, leaders of private industry, governance, and academia. They would be responsible for selecting a small group of promising teams, seeding their work with initial funding, and selecting recipients for further funds and incubator resources. Such competitions would then allow member states and private sector innovators to collaborate on bespoke solutions.
NATO’s decision to press forward with the development of AI and autonomous systems requires a complete organizational and cultural change in the alliance and among member nations. Efforts will be placed in five main policy “buckets”: standardization, responsible development, AI literacy, private sector innovation, and interoperability. And the recommendation for member states to adopt domestic technology readiness level (TRL) indicators is a surefire way to begin the shift toward impactful procurement, streamlined development, and scalable solutions.
By encouraging members to use TRLs to award, audit, and scale or terminate technologies, NATO will move closer to its end goals of interoperability, decreased overlap, strategic procurement, and streamlined R&D exchanges. Each dollar spent will thus go further, faster, and have more impact.
Nicholas Nelson – Center for European Policy Analysis, Georgia Tech Research Institute.
Nico J. Luzum – Center for European Policy Analysis, Georgia Institute of Technology.
Garritt Garcia, Anika Gouhl, Jack Sheldon, and Maria Winstead, Georgia Institute of Technology.