America is in an AI fight for its life

The U.S. Senate last week confirmed Mark Esper to serve as the next Secretary of Defense. During his confirmation hearing, Esper identified artificial intelligence (AI) as a “game changer” and the Pentagon’s most important technology. “I think artificial intelligence will likely change the character of warfare,” Esper testified, “and I believe whoever masters it first will dominate on the battlefield for many, many, many years.” The American defense establishment has awakened to this new era of AI-enabled warfare — but has much work to do and no time to waste.

It is not difficult to imagine how these AI capabilities will transform the future battlefield. The Pentagon predicts that AI will “impact every corner of the Department.” Future wars will be won or lost based on a military force’s ability to analyze and act on enormous quantities of information faster than the enemy. AI makes that possible.

China’s State Council clearly understands the importance of AI as well, issuing a 2017 strategic directive declaring that China would achieve “world-leading levels” in AI by 2030. Last week, Beijing issued a national defense strategy that lists AI as the first example of a cutting-edge technology that will lead to a new era of “informationized warfare.” 

A primary reason the Chinese Communist Party (CCP) is pursuing AI dominance is to develop the ability to coerce or defeat the United States in the ongoing great power competition and in a future crisis — perhaps without even firing a shot.

Beijing is putting its words into concerted action. While Beijing’s budget numbers are notoriously opaque and unreliable, it is estimated that China’s spending on AI will grow to at least $70 billion by 2020. In contrast, the Pentagon reportedly plans to invest about $4 billion in AI and machine learning research and development in fiscal year (FY) 2020. In fact, the Department of Defense’s science and technology budget request for FY 2020 only amounted to $14.1 billion. The Pentagon says Chinese and Russian AI “investments threaten to erode our technological and operational advantages and destabilize the free and open international order.”

For Americans, unfortunately, the reality may be even worse than this U.S. investment shortfall suggests. That is because Beijing has shrewdly exploited authoritarianism at home and theft abroad to achieve an AI advantage. 

In America, companies enjoy the right to refuse to contribute to U.S. national defense. But they also retain the right to assiduously pursue business opportunities with the CCP that the top American military officer has described as having a “direct benefit to the Chinese military.” As one might expect, Beijing does not confront a similar challenge in attempting to work with companies operating in China. In the United States, Washington asks for cooperation from private industry; in China, Beijing compels cooperation.

To achieve AI dominance, the CCP employs what the U.S. intelligence community calls a “multifaceted, long-term, whole-of-government approach” to steal or obtain the technology from the United States — including AI related technology. The U.S. Director of National Intelligence says China also uses, for example, non-traditional collectors, joint ventures, research partnerships, academic collaborations, mergers and acquisitions, and front companies to acquire foreign technology such as AI.

Chinese military officers, like their American counterparts, study military strategists like Clausewitz and Sun Tzu. “The skillful leader subdues the enemy’s troops without any fighting,” Sun Tzu famously wrote. Beijing is attempting to develop AI to make that a reality.

As U.S. military systems and the software that controls them become more advanced and complex, with potentially millions of lines of computer code, they become more vulnerable to AI-enabled cyber-attacks. The threat of software vulnerabilities is not new, but AI can revolutionize how these vulnerabilities are identified and exploited. An appropriately designed AI capability could scan for vulnerabilities across networks and platforms in seconds, and the People’s Liberation Army has thousands of individuals dedicated to identifying these vulnerabilities. If they are successful, billion dollar U.S. weapons systems may never arrive in the theater or may not function properly when they do.

In response to these challenges, the U.S. has taken positive steps. Following the issuance of the Pentagon’s AI Strategy, President Trump issued an Executive Order on Maintaining American Leadership in Artificial Intelligence. The Pentagon has stood up the Joint Artificial Intelligence Center (JAIC), and the Defense Advanced Research Projects Agency (DARPA) continues to make important progress in AI.

These efforts are positive, but to win the AI race with Beijing, Americans need to further embrace a paradigm shift. Historically, the battlefield has been dominated by analog hard power capabilities. As Secretary Esper stated in his testimony, AI-enabled capabilities will represent the most powerful and decisive weapons in future wars. Data scientists, software engineers, and cyber experts — not fighter pilots — may be the heroes of those conflicts.

This new understanding must inform every Pentagon budget request, operational concept, weapon program, recruiting strategy, and training regimen. As Christian Brose argued persuasively, “The goal should be not to buy more individual platforms but to buy faster kill chains” — and that puts a premium on AI expertise and investments.

Hard power, undoubtedly, is indispensable to American national security, but both American and Chinese hard power weapons platforms increasingly depend on software to fulfill their missions. That creates both risks and opportunities that may well determine the outcome of future wars. Consequently, considerations related to AI must be the first and last thought for every Pentagon leader and program manager overseeing a new networked weapon system enabled by software. Otherwise, that system risks becoming nothing more than an overpriced target for our adversaries.

America’s new Secretary of Defense understands the implications of AI for the ongoing great power competition and the future of warfare. Americans also need to understand that the stakes go well beyond future battlefields. The competition between China and the United States will determine what kind of international system prevails: one based on freedom or on authoritarianism. 

Lt. Col. Dan Wood is an active duty U.S. Army officer and Visiting Military Fellow at the Foundation for Defense of Democracies (FDD). He most recently studied artificial intelligence as an Army War College Fellow for Information Technology and Strategy at Carnegie Mellon University. Bradley Bowman is senior director of FDD’s Center on Military and Political Power. Follow him on Twitter @Brad_L_Bowman The views expressed are the authors’ and do not necessarily represent those of the U.S. Army or the Department of Defense.