Spy Agencies Turn to AI to Stay Ahead of Adversaries

paulista/Shutterstock.com

In a world where everyone can collect data on everything, speedy analysis could make or break U.S. national security, a top intelligence official says.

The intelligence community’s second-in-command sees technology as a primary driver of geopolitical strength in the decades ahead, and she worries the U.S. will lose its edge unless national security agencies use tools like machine-learning to accelerate their operations.

The global intelligence game has always revolved around collecting information, but the booming digital economy is making more data available to more people than ever before, according to Sue Gordon, the principal deputy director of national intelligence. In a world of “data abundance,” she said, national security will hinge on how fast countries can make sense of the information.

“The change we’re undergoing right now is as big as I’ve ever seen,” Gordon said Thursday at the Defense One Tech Summit. “This is a world where the threats are to and through information. Every technology is available to everyone, and the one that can put it to clever use faster is the one that’s going to win.”

But the traditional methods spy agencies have used to collect, analyze and share intelligence don’t necessarily fit within that new paradigm, she said. To stay ahead of global adversaries like China, Russia, Iran and North Korea, intelligence leaders are working to fundamentally change the way they make decisions based on data.

One of the main pillars of that strategy is expanding the IC’s use of artificial intelligence, Gordon said. 

Machine-learning and other data analysis tools not only bring speed to the intelligence process, she said, but the sheer quantity of available information would be nearly impossible for humans to sort through without technological assistance. To keep up with defense and intelligence agencies “insatiable” appetite for information, “we are going to have to make machines integrated into all our processes,” Gordon said. 

Spy agencies already have hundreds of AI initiatives underway, and officials are laying the groundwork to drastically ramp up those efforts in the years ahead.

The CIA is in the early stages of standing up a multi-layered cloud computing environment that would build upon the agency’s existing cloud platform, hosted by Amazon Web Services. According to contracting documents, the new system, which is expected to cost “tens of billions” of dollars to build, would provide a broad suite of analytics and storage capabilities to the entire intelligence community.

Cloud technology is likely the only way to achieve the enormous computing power required to run artificial intelligence tools at the scale of America’s defense and intelligence operations. Lt. Gen. Jack Shanahan, who leads the Pentagon’s Joint Artificial Intelligence Center, said in March the department’s AI efforts will be limited until its own enterprise cloud platform is up and running. Officials expect to award the Joint Enterprise Defense Infrastructure contract sometime in August.

Beyond expanding the intelligence community’s AI capabilities, Gordon said, officials are also working to consolidate various agencies’ data analysis and decision-making efforts. That process involves integrating disparate IT infrastructure and establishing data standards, as well as getting officials more comfortable sharing information with other members of the community.

Maintaining the country’s technological edge will also require the IC to look beyond government to industry, which makes up “90% of the nation’s threat surface,” Gordon said. 

“If [companies] are going to be the decision-makers … they have to be provided with information that will allow them to make decisions,” she said. “If they are responsible for national security, the apparatuses that service national security are going to have to figure out how we are going to also service the private sector, which we’re not used to.”

Indeed, global adversaries are increasingly setting their sights on American companies to further their own national security and economic interests. China, in particular, has for years bolstered its tech industry using intellectual property stolen from U.S. companies like Apple and IBM. China’s economic espionage tactics are at the heart of the current trade war with the U.S.

During the summit, John Demers, assistant attorney general of the Justice Department’s National Security Division, said the both U.S. tech companies and the government have a responsibility to keep China from co-opting American trade secrets for its own gain. Competition in the global tech market “will ultimately make us all better off,” he said, but IP theft could cripple the country’s innovation ecosystem.

“We’re still seeing a lot of innovation—that’s a real strength of the American system,” he said. “What we’re focused on is not having that innovation being stolen.”