Defense Department Unveils Artificial Intelligence Strategy

To maintain its strategic position in the world, succeed on future battlefields and protect the homeland, the Department of Defense must increase the adoption of artificial intelligence, according to the department’s newly released Artificial Intelligence Strategy.

“Artificial intelligence (AI) is rapidly changing a wide range of businesses and industries,” the document stated. “It is also poised to change the character of the future battlefield and the pace of threats we must face. We will harness the potential of AI to transform all functions of the department positively, thereby supporting and protecting U.S. service members, safeguarding U.S. citizens, defending allies and partners, and improving the affordability, effectiveness and speed of our operations.”

The strategy singles out China and Russia as making significant investments in militarized AI, including applications that cast doubt on the adherence to “international norms and human rights,” the document noted.

In contrast, the DOD promised to lead in the responsible use and development of AI, stating, “The department will articulate its vision and guiding principles for using AI in a lawful and ethical manner to promote our values.”

The DOD intends to use AI to support and protect U.S. warfighters and civilians globally as well as safeguard the country and U.S. citizens at home. As an organization, the DOD will use AI to become more efficient and streamlined, refining workflows and improving the speed at which tasks are performed. The department also aims to be an innovator in scaling AI across its global operations through the use of a common platform “that enables decentralized development and experimentation,” according to the document.

In particular, the military will apply AI capabilities to support key missions—at a rapid and iterative, yet responsible way. Situational awareness, decision making, equipment operational safety, predictive maintenance and supply, and business processes can all benefit from AI-enabled tools, the document stressed.

“We will prioritize the fielding of AI systems that augment the capabilities of our personnel by offloading tedious cognitive or physical tasks and introducing new ways of working,” the strategy dictated.

At the heart of the DOD’s AI efforts will be the Joint Artificial Intelligence Center, known as JAIC, and pronounced like Jake. DOD Chief Information Officer Dana Deasy’s office established the center in June 2018 to lead AI-related planning, policy, oversight, synchronization, ethics and safety within the department.

JAIC is initially focusing on near-term execution and AI adoption, the strategy stated. JAIC also will harness the long-term AI-related efforts by the Defense Advanced Research Projects Agency (DARPA), DOD laboratories and other entities such as the Defense Innovation Unit (DIU).

Using cross-functional teams (CFTs) across the department on a rotational basis, JAIC will identify national mission initiatives (NMIs) to prototype and deploy AI capabilities to the Joint Force. For more specific projects, JAIC will help address AI deployments for so-called component mission initiatives (CMIs).

“Both NMI- and CMI-type efforts will include selecting commercial and academic partners for prototypes and employing standardized processes with respect to areas such as data, testing and evaluation, and cybersecurity,” the strategy stated.

In addition, according to the strategy, the DOD will turn to industry, academia and international allies to form “strong partnerships…. at every stage in the AI technology pipeline, from research to deployment and sustainment.”

Speakers last week at AFCEA’s Rocky Mountain Cyberspace Symposium 2019 attested to the importance of AI for the DOD.

“I’m pleased to see the department move forward in this regard,” said Gen. Dennis Via, USA (Ret.), fellow, Defense Futures; and senior executive advisor, Booz Allen Hamilton.

Gen. Via noted that the John S. McCain National Defense Authorization Act (NDAA) for fiscal year 2019 “was very explicit” in addressing the development of AI technology as it relates to national security. In particular, the department can benefit from the capabilities AI brings to data. “On the defense and intelligence front, agencies have long sought to take full advantage of the most valuable resource, the vast amount of data collected on a hourly and daily basis,” he stated.

“They want to be able to use that data to make more insightful, forward-looking decisions about readiness, logistics, manpower, intelligence and a host of other critical national security concerns.”

In addition, Vice Adm. Nancy Norton, USN, director of the Defense Information Systems Agency (DISA) and commander, Joint Force Headquarters–Department of Defense Information Network (JFHQ-DODIN), confirmed that DISA was already working with JAIC to implement AI initiatives, as well as applying AI technologies to DISA’s existing portfolio through a DISA-led team.

Meanwhile, the Air Force Academy, which spent two years revamping its cyber curriculum, is now looking at how to incorporate AI and machine learning into its program, according to Lt. Col. Mike Chiaramonte, USAF, associate professor of computer science, U.S. Air Force Academy; and director, Air Force CyberWorx.   

“At the Air Force Academy, we are looking at how do [we] build future military leaders that can operate in that domain, in a world where human-computer teams are much more deep and rich than they are today, and where machines are perhaps making decisions that humans are making today,” he stated. “When we start bringing in machines into our warfighting teams as decision makers, the dynamics change a bit. Ethics come into play as does trust, and how do I trust a computer team member.” 

Implementing AI within the DOD comes down to more than the math needed to develop the algorithms, Col. Chiaramonte stressed. It involves developing the right policies, laws and rules of engagements, he said.

To Gordon Hannah, principal, Risk and Financial Advisory at Deloitte, the near- and mid-term implementation of AI and machine learning will result in shifting demands on the cyber workforce. In order for the DOD—and others—to implement AI and machine learning, they need data scientists, mathematicians, sophisticated analysts that understand network connections, better user interfaces and more infrastructure specialists—particularly in cloud infrastructure—as well as advanced developers who can deploy machine learning and AI.

“We think the best approach is for AI to augment humans, not replace them, especially in cybersecurity,” he noted.