Can Artificial Intelligence Compensate for Strategic Shortcomings?

To frame this question, we should first note that the United States has not won a war in more than 20 years (if we count Serbia as a "win"). Nor has it has had an effective strategy. This is not a criticism of the military, but of its civilian leadership.

Second, artificial intelligence (AI) is not good at developing strategy. Perhaps this will change as the technology matures, but we cannot expect AI alone to remedy our current weaknesses. AI is a vastly improved computing tool, but it cannot conceptualize and combine interests, goals, and means in ways that develop strategy. Strategy development remains a human function.

Third, better technology does not guarantee success. Building the fastest car and giving it to a cautious driver is unlikely to win a race, particularly against skilled competitors. (How skillful our competitors are is a different question, but in key regional competitions in Europe, Asia, and the Middle East, they have been more skillful than the United States at obtaining outcomes that advance their interests.) The experience in Afghanistan shows that enormously superior technology and exceptional forces, when married to impracticable strategy, do not produce success even against a primitively equipped but determined opponent.

The military benefits of AI are in themselves insufficient to offset reverses that result from misguided strategy. Weapons may eventually become faster and more precise, sensors more capable, logistics streamlined, and the compilation of masses of intelligence may be improved, but this does not compensate for misdirection. Incremental improvements can be beneficial—but not beneficial enough to give the United States a clear advantage in the current contests with hostile powers.

The notion of marginal improvement runs against much of the received wisdom on emerging technologies, which include assertions that that AI will change everything. This comes after being told that the Cloud will change everything, also blockchain, 5G, or whatever the technology du jour happens to be —all will change everything. Change will happen—but only after making the adjustments needed for new technology to provide economic or military benefit. The last technology to change everything was the internet. The internet was running in the 1980s, was commercialized in the 1990s, but took at least a decade more for it to begin to have a substantial effect. AI, which has also been around for decades, can best be seen as another phase in the evolution driven by software and digital networks and will follow a similar trajectory in its effect.

As part of any strategic rethink to adjust to AI, the United States will need to reconsider some of the major weapons systems it has used for decades. Aircraft carriers are the weapon that won the naval battles of World War II, but their utility is threatened by hypersonic missiles. Ballistic missile submarines entered service in 1959, but AI may eventually make it easier to discover them in their deep sanctuaries. While these improvements in detection remain unproven, they must be considered in planning and acquisition. Work on stealth aircraft began in the 1970s, but we now face the end of stealth. Our opponents have been working to defeat stealth for 20 years, and stealth probably has another decade of utility as air defense sensors and processing continue to improve. Simply layering AI over older weapons and old strategies won't change these trends.

Some years ago, "cyber" was treated as magic technological fairy dust that could be sprinkled over military problems (at least in exercises) to instantly and painlessly resolve them. This did not work in real life, and it took years of planning, exercise, visionary leadership, and a little experience to turn cyber operations into a valuable military tool. The same will be true for AI applications (many of which are at best prototypes). This will not be another "Revolution in Military Affairs" until the concepts behind military operations are rethought, and weapons that use new technologies are deployed. At an operational level, this means exercises and experimentation, but new technology will not provide military advantage until it is used to support better strategies.

In assessing the effect of the current generation of new technologies, bear in mind that at the dawn of the internet, there were many examples of major companies that led technology development. These companies went out of business when they attempted to just layer the new technology over their existing business models, or when they failed to recognize that business models developed by new firms not burdened with legacy strategy made these companies more profitable.

Technologies like AI require significant investment and change before they can generate results. New technologies can prop up old strategies, but only until those old strategies are challenged and fall over in defeat. The United States retains innate advantages over its opponents, and there is a growing recognition of the deficiencies of strategy; but looking at earlier generations of tech giants, staying competitive means we need new strategies more than new technology.

James Andrew Lewis is a senior vice president and director of the Technology Policy Program at the Center for Strategic and International Studies in Washington, D.C.

Commentary is produced by the Center for Strategic and International Studies (CSIS), a private, tax-exempt institution focusing on international public policy issues. Its research is nonpartisan and nonproprietary. CSIS does not take specific policy positions. Accordingly, all views, positions, and conclusions expressed in this publication should be understood to be solely those of the author(s).

© 2020 by the Center for Strategic and International Studies. All rights reserved.

James Andrew Lewis
Senior Vice President; Pritzker Chair; and Director, Strategic Technologies Program