By ADM James Stavridis, USN (Ret.)
Originally published in Bloomberg – May 20 2021
With the release of the much-anticipated National Security Commission on Artificial Intelligence report, the U.S. must confront an inconvenient truth: America, in the words of co-chairmen Eric Schmidt and Bob Work, “is not prepared to defend or compete in the AI era.” Schmidt, the former chief executive of Google, and Work, former deputy secretary of defense, are as deeply versed in this subject as anyone in government or the private sector. Americans should treat this threat as a looming tower. What is the state of play and where does the U.S. go from here?
First, let’s address the most obvious and concerning opponent in the AI field: China. The Chinese have been focused relentlessly for a decade on AI, and have built a “whole of government approach” from the ground up, to become the dominant global AI power. This includes a massive focus on science, technology, engineering and math education at every level; grooming of the brightest prodigies in the field; increasing levels of research and development, including for military operations; garnering control of vast pools of data from social networks (data is the “new oil,” as the saying goes); industrial espionage to gather solutions from international competitors; and incorporation of AI into every field of government and private endeavor. It is a clever and consistently applied strategy to dominate this vital field.
Meanwhile, Washington has primarily adopted a market approach, on the hope that Silicon Valley and other tech centers will fund much basic research and keep America in the game. The U.S. has lagged in STEM education, and privacy rules preclude much gathering of massive data sets for experimentation. While some U.S. government and university labs are engaged in AI research (I am a senior fellow at the Johns Hopkins Applied Physics Lab, a leader in the field), and the military is contracting some of this work, the approach is thus far too laissez-faire. There are some things — putting a person on the moon in the 1960s comes to mind — that the private sector won’t initially have the incentive to do on its own. AI fits in that category.
Most importantly, AI is important in many different fields, from manufacturing to advertising to medical treatment to agriculture. But the zone in which the greatest danger of lagging behind principal opponents is quite clear: military operations.
The uses of AI in military activity are both broad and deep. As a Navy and NATO commander, I faced hard decisions in combat — during operations Desert Shield and Desert Storm, and in the wars in Afghanistan, Iraq and Libya — that had to be executed with limited knowledge and no way to compare my choice with how the other options would turn out.
An AI adviser would change that. Like someone who has played a trillion games of the Chinese strategy game Go, a combat commander’s AI adviser could use computing power to synthesize information and intelligence, order and prioritize its presentation to the decision-maker, and ultimately provide advice on next moves.
This applies in military operations from the level of a SEAL team petty officer clearing a terrorist house up to the decisions I made as a four-star admiral in Afghanistan. The applications are powerful and will significantly elevate the performance of the military that gets there first.
The U.S. will need a coherent, all-of-government strategy that is aligned with the private sector. Some of the key elements, many of which are identified in the AI report, include investment in research and development. The report recommends a $40 billion down payment; identification of talent pools with incentives to bring the best into government service; a national academy for cyber; a Defense Department Cyber Force with deep expertise in AI (akin to and interwoven with the new Space Force); top-level advice to the president and the National Security Council; and more aggressive recruiting of top international talent through immigration.
An example of how the private sector fits in is the stark geopolitical fact that so many top-end microchips are manufactured in Taiwan, which is under constant pressure from China. The U.S. must look at global supply chains and onshore the ones that make sense, starting with microchips.
America’s lead in AI is tenuous and shrinking, and some observers believe it is already gone. Artificial intelligence is inextricably tied to cyber capabilities, both offensive and defensive. Especially with the advent of quantum computing (a vastly more complex structure advancing from simple binary bits to moving information by positioning a wide range of particles in different positions), the entire set of premises upon which we have based cryptology, computer security and command-and-control networks are increasingly outdated.
The military operators in all of the services, and especially the intelligence community, must be part of embracing the AI revolution. It is, unfortunately, the battle America is least prepared to fight.
To contact the author of this story:
James Stavridis at firstname.lastname@example.org