When people think about what military applications of artificial intelligence (AI) will look like in the future, they tend to imagine killer robots.
Michael Horowitz explains that, in reality, most of the important military uses of AI will not be on the battlefield. Instead, these AI applications will be in areas such as logistics, personnel and other back-office areas of the military. One potential outcome of the increasing military use of AI is a transition in the military’s balance of power. The combination of technological change and organizational adaptation necessary to take advantage of AI can be risky for great powers, such as the United States, that may rest on their laurels, rather than take chances. Rising powers, such as China, may be more willing to take risks to benefit from AI, which could give them an advantage in thinking about what military integration of AI looks like.
A shift in the global balance of power would present a significant risk to global stability, therefore, there is a need for global governance solutions to mitigate potential harms.
Since AI is a general-purpose technology, with applications mostly in the civilian domain but also in the military domain, there remains a lot of uncertainty about the role AI will play in shaping military power. This often happens during the emerging periods of new technologies. A classic example of this in world history comes from the aircraft carrier. The British invented the aircraft carrier at the end of World War I. But they thought about the aircraft carrier and aircraft as spotters for their battleship, because they were the best in the world at battleship warfare, and they were organizationally constrained. In contrast, the United States and the Japanese — trying to figure out how to project power over the vastness of the Pacific — used aircraft carriers as mobile airfields and unlocked that potential in a way that transformed naval warfare.
Today, when people think about what military applications of artificial intelligence will look like they tend to imagine killer robots. But, certainly in the near term, most of the important uses of artificial intelligence by militaries will not be on the battlefield. Instead, they’ll be in areas like logistics, personnel and other back-office areas of the military. And one potential outcome of an increasing use of AI by militaries is a transition in the military balance of power. The combination of technological change and organizational adaptation necessary to take advantage of it can be really risky for great powers, like the United States, that might rest on their laurels rather than taking chances. Whereas rising powers, like China or others, might be more willing to take risks to benefit from the potential of artificial intelligence — that could give them an advantage in thinking about what AI integration for militaries looks like.
The increasing integration of artificial intelligence into militaries around the world is probably inevitable. But how it happens and what it means is still very up in the air. The potential risks for international stability from artificial intelligence have generated a lot of interest in potential global governance solutions as a way to mitigate potential harms. And, in particular, confidence-building measures — like trust, notification, transparency — will be critical in the years ahead, as countries try to understand what military applications of AI will mean for the future of power.