AI Will Change War, But Not in the Way You Think

September 2, 2019

This article was submitted in response to the call for ideas issued by the co-chairs of the National Security Commission on Artificial Intelligence, Eric Schmidt and Robert Work. It addresses the first question (part a.) which asks how artificial intelligence will change the nature of war.

The Terminator? The Marines’ idea of a perfect rifleman. Neo in The Matrix? The Army’s concept for picture-perfect, human-machine symbiosis. R2-D2 from Star Wars? The Air Force’s dream of a loyal wingman.

However much the military might wish to take these technologies from the screen to the battlefield, the reality of artificial intelligence (AI) is far less sexy than Hollywood would have you believe. Unlike science fiction movies, which depict the technology itself as an instrument of war, AI will function primarily as an enabler. AI will change how wars are fought, but not the nature of war. War is still, and will forever be, applied violence to achieve a political goal. Like human nature itself, the fundamental qualities that define the nature of warfare are impervious to change. AI will, however, cause a shift in the character of war.

Nevertheless, AI is sure to permeate every aspect of warfighting — from movement to communication, logistics, intelligence, weapons, and people. Delivering these warfare-changing technologies to the frontlines and into American hands will depend on less glamorous activities — namely, expediting the procurement process to more quickly field AI, and securing supply chains by collaborating with U.S. companies.

The Need for Speed in Acquiring AI

The race to acquire AI will be different than the race for the atomic bomb, intercontinental missiles, or precision-guided munitions. Some aspects will be similar — nations will attempt to be the first to develop and acquire AI systems and applications. But rather than a single space race “Sputnik moment,” there will be continuous milestones in the AI race — algorithm updates, software patches, etc. The near-term difference between AI and previous technological innovations is that AI is iterative, incremental, and, most importantly, an enabler of all parts of warfare. Moore’s Law states that the processing capability of computers can be expected to double every two years. Due to this iterative cycle and the continued breakthrough in AI abilities and applications, the speed at which the warfighter receives the technology is paramount. Now, megabytes may matter more than megatons.

For example, for his book Army of None, Paul Scharre interviewed former Undersecretary of Defense for Acquisition, Technology, and Logistics Frank Kendall. Kendall states, “Automation and artificial intelligence are one of the areas where the commercial developments I think dwarf the military investments in R&D. They’re creating capabilities that can easily be picked up and applied for military purposes.” Kendall had it almost right.

It’s true the commercial sector is indeed dwarfing military investments in R&D. As MarketWatch reported, “In 2015, for example, the combined research and development spending at the U.S.-headquartered companies Google, Apple, Facebook, IBM, Microsoft and Amazon was $54 billion.” By contrast, Defense Department R&D into AI doesn’t even come close. According to Bloomberg Government, “the Defense Department plans to allocate $4 billion toward artificial intelligence and machine learning R&D activities in fiscal 2020” — startlingly only 0.5 percent of the departmental budget and less than one-tenth that of the six companies mentioned before.

What Kendall overstated, however, is that the United States cannot easily acquire these capabilities today. The Section 809 Panel, which the Fiscal Year 2016 National Defense Authorization Act established to “streamline acquisition regulations,” focused on how the Defense Department makes purchases — everything from paperclips to Patriot missile batteries. As the panel stated in its summary of recommendations from Volume 3 of its final report issued January 2019, the Pentagon’s “slower acquisition pace has a direct effect on warfighting capability in a defense era defined by technological edge … DoD must acknowledge its acquisition system suffers from processes and procedures that are obsolete, redundant, or unnecessary and work to move quickly enough to keep pace with private-sector innovation.” Recognizing the problem, the panel advocated revolutionary and evolutionary procurement practices that are fit for a 21st-century competition. For example, recommendation 81 clarifies and expands the use of Other Transaction Authority (OTA) for rapid government utilization of technology. It’s now up to Congress and the Pentagon to implement this recommendation and others the panel suggested.

If the United States does not fundamentally shift its thinking on how to adopt commercially-available and government-unique AI systems, it will surely see a decline of hard and soft power as the world adapts and other countries don’t play by the United States’ same arcane rules. 

Securing Supply Chains

Securing the government’s procurement pipeline is the next logical step in enabling widespread adoption of artificial intelligence. In Peter Singer and August Cole’s novel Ghost Fleet, which presents a hypothetical war against China, the F-35 is compromised by malware implanted in computer chips on the aircraft. This allows China to activate a kill switch and instantly take F-35s out of the fight. This scenario is not just the stuff of fiction — in the Super Micro supply chain compromise, Chinese intelligence services had a manufacturer plant malicious chips on hardware exported to the United States.

To achieve the enabling power of AI, domestic technology companies should partner with the very entity that has driven their success — the U.S. government. Likewise, the government should recognize that technology companies have caused a shift in how technological advances are made, leaving Silicon Valley, not Uncle Sam, the driver of research and development. In order to accomplish this, the culture of government ought to evolve to embrace that shift. To broker the trust needed, technology companies should take ownership of the fact that their products, if in the hands of America’s enemies, can cause the nation great harm.

Just like the fictitious company Cyberdyne developed and produced the Terminator, industry, not government, will largely drive the warfare-disrupting AI technologies of the future. In the not-so-distant future, an AI-enabled logistics system could track fuel usage trends of a surveillance unmanned aerial vehicle. Autonomously, this system could realize the unmanned aerial vehicle will need fuel. The system would be able to incorporate AI-enabled intelligence to determine the optimal driving route that delivers the unmanned refueling vehicle safely to the airbase in need. An AI-enabled robot could then autonomously refuel the unmanned aerial vehicle and communicate back to the mission owner that its job is complete. Not exactly the Terminator raining down death and destruction.

A domestic private sector that stands ready to work cooperatively with the government is the linchpin to securing the artificial intelligence supply chain. The first American company to invest R&D dollars heavily in purely domestic AI development and manufacturing will cause a paradigm shift in technological advances and instigate other companies to follow suit. In fact, companies like Palantir and Anduril are starting to advocate just that. Will it cost more? Sure. But the cost of the alternative is far greater and far darker.

In the past, U.S. firms were more than willing to work in defense of the country. Ford Motor Company supported American manufacturing efforts in World War II by building engines, planes, and tanks, which helped turn the tide against the Axis powers. While the company’s war record was not perfect, Ford recognized its fate was linked with that of the nation. The United States incentivized the company’s success through trade policies, infrastructure investment, and other economic stimuli,  and Ford rightly reciprocated when faced with the threat to the American way of life. So, too, should American technology companies today.

American firms like Google, Apple, Facebook, Microsoft, and Amazon have driven considerable technological innovation and employed thousands. Americans, immigrants, and the nation’s economic policies fueled the meteoric rise of these companies, yet disturbing trends have emerged. Namely, Google’s announcement that it would not renew its contract with the Department of Defense’s Project Maven, an AI program meant to analyze unmanned aerial vehicle surveillance footage. The cry from some of Google’s workforce — the same cry that prompted its executives not to renew — was that Project Maven itself is a weapon of war. Ironically, this AI application is meant to detect enemies and will provide military targeteers with higher fidelity battlefield information that can spare innocent lives from collateral damage. Effective implementation of the program could save lives.

To be fair, trust is a two-way street, and Washington has broken trust with industry several times in recent years. Recent examples include the growing Joint Enterprise Defense Infrastructure (JEDI) cloud computing contract blunder, worth up to $10 billion, and the Apple-FBI debate over encryption of Apple products following the San Bernardino, California shootings in 2015.

In the JEDI example, the Pentagon said it wanted to move fast, heeding the call from industry to quickly adopt cloud capabilities — utilizing the same OTA the Section 809 Panel recommended. Last year, the Defense Department released a draft request for proposal. The government has made little progress since in keeping pace with technological change. Government officials from President Donald Trump to congressional representatives are calling into question the process behind the pending JEDI award to select a vendor for the Defense Department’s cloud infrastructure over the next ten years. American companies like Amazon, Microsoft, Oracle, and others stood ready to do business with the federal government. Instead, the government got in its own way, dramatically slowing the speed at which the Defense Department can acquire the envisioned capabilities. The process is now at a crawl with Secretary of Defense Mark Esper promising a full review of the proposed contract just weeks before the Department was scheduled to announce a winner. The problem with JEDI wasn’t the speed at which the process moved — the problem was a parochially-minded and antiquated system not fit for 21st century warfare.

Rebuilding trust between the government and the private sector is necessary to facilitate the incorporation of AI into the American arsenal — that will enable a dramatic shift in the character of war. Regrettably, the United States stands unprepared today to adopt AI for the wars of tomorrow.

Upsetting the Status Quo

If the government can grow more imaginative and shed self-imposed bureaucratic constraints, envision the art of the possible by partnering with industry, and lay the foundations to ensure delivery of a key enabler for future warfare, the United States will be well positioned to advance its interests in the 21st century — ahead of China and Russia.

To get there, the United States doesn’t need a Manhattan project. It needs another Boston Tea Party. Throw the old way of doing business overboard and embrace the tech sector driving AI R&D. The status quo should be upset and a few dangerous thinkers, in government and business, empowered to be revolutionary. Industry and government together should look introspectively and appreciate that their successes are mutual and interdependent. American industry should recognize that its success depends on the country’s success, and the government should act smarter and faster to keep the country safe.


Jonathan Clifford currently serves as a consultant to the federal government and an officer in the U.S. Navy Reserve. He previously served as a national security guru for a technology trade association and as a senior advisor to members of Congress in the House of Representatives, specializing in national security and the members’ work on the Armed Services and Intelligence committees. The views above are Jonathan’s and are not reflective of any employer. For more #natsec and rarely funny tweets, follow him on Twitter @JClifford33.

Image: U.S. Army