Artificial Intelligence Will Put Spies Out of Work, Too
Secret mapping agency aims to automate the bulk of its work.
If Robert Cardillo has his way, robots will perform 75 percent of the tasks currently done by American intelligence analysts who collect, analyze, and interpret images beamed from drones, satellites, and other feeds around the globe.
Cardillo, the director of the National Geospatial-Intelligence Agency, known by the acronym NGA, announced his push toward “automation” and artificial intelligence at a conference this week in San Antonio. The annual conference, hosted by the United States Geospatial Intelligence Foundation, brings together technologists, soldiers, and intelligence professionals to discuss national security threats, changes in technology, and data collection and processing.
Artificial intelligence is on the rise; former President Barack Obama’s White House released a white paper on its potential future impacts in the final months of the administration. Police officers are using preliminary programs to predict the likelihood someone will commit a crime in a specific neighborhood based on crime statistics data. And companies like Amazon and Netflix use machine learning to calculate what movie you will want to watch or which book you may buy.
Yet this sort of automation is also seen as a threat to workers, who fear being put out of jobs, particularly in the private sector.
The fear that artificial intelligence will take over jobs, or fail catastrophically along the way, is palpable in the intelligence community as well, and Cardillo admitted that the workforce is “skeptical,” if not “cynical” or “downright mad,” about the prospect of automation intruding on their day-to-day lives, potentially replacing them.
The coming revolution in artificial intelligence has been hyped for years, often falling short of expectations. But if it does happen, analysts worry they’ll become obsolete.
Cardillo, who called it a “transforming opportunity for the profession,” said he’s working on showing the workforce that artificial intelligence is “not all smoke and mirrors.” The message he’s sending to workers at the agency is that the goal of automation “isn’t to get rid of you — it’s there to elevate you.… It’s about giving you a higher-level role to do the harder things.”
In Cardillo’s eyes, the profession of geospatial intelligence — monitoring and exploiting commercial and proprietary video and imagery feeds around the world — is on the precipice of a data explosion similar to when the internet took off. At that point, the National Security Agency, which is responsible for collecting and analyzing digital communications, had to figure out ways to vacuum up and glean specific conclusions from an explosion of communications traveling back and forth on the web.
Just as the NSA employs algorithms to trawl through millions of messages, Cardillo wants machine learning to help with large volumes of imagery. Instead of analysts staring at millions of images of coastlines and beachfronts, computers could digitally pore over images, calculating baselines for elevation and other features of the landscape. NGA’s goal is to establish a “pattern of life” for the surfaces of the Earth to be able to detect when that pattern changes, rather than looking for specific people or objects.
NGA is responsible for tracking potential threats, such as military testing sites in North Korea. When something at a site changes, like large groups of people or cars arriving, it may indicate preparations for a missile test. “We don’t have a higher priority,” Cardillo told Foreign Policy. “We put everything we can into North Korea.”
But the number of sensors, images, and video feeds is exploding and will continue to grow in the coming years, he predicted. “A significant chunk of the time, I will send [my employees] to a dark room to look at TV monitors to do national security essential work,” Cardillo told reporters. “But boy is it inefficient.”
The agency is also turning to academia and the private sector for help. Cardillo hired Anthony Vinci, the founder and former CEO of Findyr, a company that crowdsources data from countries around the world, to head up the agency’s machine-learning efforts within NGA.
Companies exhibiting at the conference were clearly on the artificial bandwagon, boasting flashy datasets and advanced algorithms. But not everyone was convinced relying on computers for the bulk of data crunching and analysis was such a great idea for intelligence work.
Justin Cleveland, a former intelligence official who works for the security company Authentic8 — which created a secure browser called Silo that also allows intelligence professionals to disguise their cybertracks — was skeptical of the automation boom. “It can be helpful,” he said in an interview at the conference. “But you could have one bad algorithm and you’re at war.”
Taking humans out of the bulk of the process is bound to lead to errors. “At the end of the day, you have to trust the person who wrote the algorithm over the analyst,” Cleveland said.
Jimmy Comfort, a deputy director at the National Reconnaissance Office, was enthusiastic about certain applications for artificial intelligence in some areas like facial recognition. “There are so many parallels with what the commercial guys are doing,” he said in an interview.
But for his agency, which works mainly with satellites, the needs are different. Satellites take fewer images, from much farther away. “There’s challenges for us doing that stuff from space,” Comfort said.
Photo credit: CHIP SOMODEVILLA/Getty Images