Key Takeaways for Executives
- More and more federal agencies are starting to use artificial intelligence to accomplish their missions.
- This is especially true in the national security area, where analysts employ AI to help them sift through torrents of intelligence data.
- AI complements the work of human analysts, who contextualize and interpret the data processed by AI.
- The Defense Department and other agencies are looking to the private sector, particularly Silicon Valley firms and other “non-traditional” contractors, to help develop AI solutions.
The phrase “artificial intelligence” conjures up all sorts of images, mostly from movies featuring sentient robots bent on conquering humanity.
Thankfully, AI in the real world is much more limited and designed to assist humans rather than enslave them. And contractors are looking to take advantage of the government’s growing interest in this burgeoning field.
Balan Ayyar, CEO and co-founder of Percipient.ai, said currently, AI solutions for the government are used to support “narrowly tailored artificial intelligence for very specific mission sets and problems.”
Ayyar said analysts and others in the national security community, although extremely well-trained, are overwhelmed with torrents of data from drones, sensors, cameras, internet-connected devices, and so forth.
For example, the number of connected devices in use worldwide has been estimated at more than 17 billion in 2018 — not including smartphones, tablets and laptops — and that number is growing exponentially.
“The enemy can hide in plain sight,” he said. “There’s a recognized crisis in national security that is driving adoption of AI.”
AI is unable to contextualize any of this information or understand what it means. A computer can’t tell whether someone with a knife is a surgeon or a violent criminal, or whether someone with a blue backpack in a video image is a college student or a terrorist also known to be carrying a blue backpack.
But agencies can use AI to identify trends, anomalies and other items of interest. In other words, Ayyar said, the goal of AI is not to create a sentient machine — such an accomplishment isn’t even on the horizon — but to help human analysts sift through giant piles of data that grow larger every day so they can do what they do best: determine what is relevant and act on it.
The government recognizes AI’s potential, and federal agencies are slowly beginning to adopt its capabilities.
“For the first time, they’re not being trapped by outdated systems,” Ayyar said. Instead, agencies are looking for integrated operations at scale and to “keep control of their data. They now understand the power of their data and recognizes there’s much more data coming in than they can process. They can’t hire enough people to solve this problem.”
For example, last year the DOD chief information officer created Joint Artificial Intelligence Center, which will oversee most of DOD’s AI initiatives. JAIC will develop a set of AI “standards…. tools, shared data, reusable technology, processes, and expertise” for the national security community, according to a memo.
JAIC’s first successful use of AI came in September 2018 in response to Hurricane Florence. During the aftermath, JAIC, along with the Federal Emergency Management Agency, U.S. Customs and Border Protection, the National Guard and the Coast Guard, developed a tool using AI that helped first responders map victims in need of rescue.
DOD is looking to private software firms in Silicon Valley and elsewhere to develop solutions for AI and other needs, Ayyar said.
The private sector needs to step up, he said, because adversaries like China and Russia are going all-in on AI. For example, China has a national strategy designed to make it a world leader in AI by 2030.
Even now, Ayyar said, there’s a “grave concern” both China and Russia are starting to use AI in nefarious ways, for example, by interfering with the U.S. electoral process through Facebook and other social media platforms.
“We’re at a Sputnik-like moment here,” he said. “Contractors should understand this trend line and how they can help.”
Related: Top AI Execs to Watch