MEDIA

HARNESSING THE POWER OF ARTIFICIAL INTELLIGENCE

By Kim Catley

The next generation of leaders will be fluent in the language of artificial intelligence (AI). Today, U.S. Naval Academy graduates are using AI to support ocean exploration, prevent mass shootings and improve defense capabilities.
As officers in the fleet and in the Marine Corps, Naval Academy graduates are charged with making decisions that matter. Part of their training at the Academy is aimed at using evidence and data to make smarter decisions—a process that is now made even better with the help of artificial intelligence and machine learning.

“AI and machine learning can include a very large amount of data from many sources and summarize that information into a recommendation,” said Gavin Taylor, professor of computer science at the Naval Academy. Taylor is also the co-director of the Academy’s Center for High Performance Computing Education and Research, which was founded after the Department of Defense provided advanced technology with the hope midshipmen would graduate with more experience in the high-performance systems.
The center aims to complement academic programs—like the new data science major and the computer science certificate in artificial intelligence—with funding for internships, research, conferences and other experiential opportunities. It also provides seed grants for faculty members to integrate performance computing and machine learning in their own research.
“We have all these people who are going to graduate and be at the sharp end of these DoD-specific problems,” Taylor said. “The idea was to get the Academy involved early so that when the students graduated and went out in the fleet, they would have accounts on these big computers already and could start trying to solve these problems.”

However, he says, the Academy is more than a training center focused on specific systems. Graduates need a deeper background that equips them to learn about and use new technologies as they’re developed.

That foundation is evident in several alumni businesses that integrate AI and machine learning—even though the technology has evolved rapidly since they were midshipmen. Here, they share why AI is integral to their businesses in the private sector, and how they’re contributing to the future of security and defense.

‘WE’RE PROBLEM SOLVERS’
The story of ZeroEyes begins in Parkland, FL, following the 2018 mass shooting at Marjory Stoneman Douglas High School. ZeroEye’s founders, all former Navy SEALS and military veterans, had lived through active shooter situations and wanted to find a proactive way to keep people safe.

As they brainstormed, they kept coming back to one common element of most mass shootings: the perpetrator could often be found holding a rifle in view of an active security camera.
“The vast majority of active shooters start in a staging area—like parking garages, loading docks or stairwells—that are covered with existing cameras,” said Kieran Carroll ’07, the company’s chief strategy officer. “The problem is no one has the ability to watch hundreds or thousands of cameras.

“But we’re problem-solvers. We started looking at a technology solution that could mitigate, or hopefully prevent, the next shooting from happening.”

They developed a software solution that integrates with the existing cameras and uses AI to detect a visible firearm. An alarm sounds and the staff at ZeroEyes’ operation center make a rapid assessment of the situation and can contact the authorities and first responders when warranted.

The software-to-human hand-off is critical to ZeroEyes’ success, said Carter Stapleton ’07, vice president of operations. While humans don’t have the capacity to carefully monitor hundreds of cameras for hours on end, AI can’t always accurately detect a weapon and false positives can cause a major disturbance. ZeroEyes also intentionally stacks the operation center with military veterans who are trained to identify a weapon and conduct a threat assessment in three seconds or less.
“Our goal is always to de-escalate,” Carroll said. “That’s why we hired the military analysts who have the right constitution to make these critical decisions quickly, and with an absence of wider information. One analyst can manage thousands of cameras simultaneously without getting overly saturated.”

While education settings are the core of the business, ZeroEyes is further expanding into commercial and government settings. Julia Fries ’19, a sales project manager who is exploring DoD innovation efforts around AI and machine learning, said the ZeroEyes system can provide base security and support for combat environments.
“We have ways to mitigate cognitive overload for operators, and use AI for greater situational awareness involving surveillance and intelligence,” she said. “It’s different from the schools, but still part of our focus on providing force protection.”

‘HOW DO WE MAKE ROBOTS SMARTER?’
At Terradepth, CEO Joe Wolfel ’04 is using unmanned robots to explore and map an area humans know little about: the vast floor of the ocean.

Using acoustic data, sonar, laser line scanners, optical cameras and more, Terradepth’s robotic fleet is creating a high-resolution geophysical data map of the seabed. The data can then be used to determine where to place an offshore wind farm or submarine telecommunications cable, or studied over time to evaluate the effects of climate change.

Only 5 percent of the ocean floor has been precisely mapped—meaning the potential data to be collected is immense. To make the process scalable, Wolfel and his team are turning to AI. They’re focused on two avenues in particular: data processing and initial analysis by the robots, and delivering energy to sensors without human intervention.

“How do we make robots smarter, and how do we make them stay out longer?” Wolfel said. “Because humans are the limiting factor, from an exposure risk and a cost standpoint.

“But the question isn’t how do robots fully replace you, but how do they minimize human cognitive load or involvement? What you don’t want is to set a robot loose for two weeks, come back, and realize the algorithm failed.”

Building a scalable fleet of smart robots is an initial priority, but Wolfel said AI and machine learning will also be critical as they develop systems to automatically detect changes in the seabed and, eventually, create augmented reality fly-through environments that allow users to experience the ocean floor.

Wolfel, who also served as a Navy SEAL, says he realized the value of AI from the outset, but he had little experience with the technology before launching Terradepth. However, he said, his broad-based education at the Academy gave him the necessary exposure to complex, cutting-edge technology that allowed him to hire the right experts and lead with awareness.

‘CONTRIBUTING TO THE GREATER DEFENSE MISSION’
After a career that took him from supporting Naval Special Warfare on the West Coast to working for the Joint Chiefs of Staff, the Senate and the White House, Andrew Keene ’12 shifted to the private sector in early 2020 when he joined the defense technology startup Anduril Industries.

The 5-year-old company specializes in cutting-edge autonomous defense capabilities, such as surveillance systems to detect drone threats and protect military bases and enforce borders. It also produces advanced command-and-control software as well as autonomous air and undersea vehicles used for a variety of missions. Everything leverages a proprietary AI-powered software, Lattice OS, which is the core of Anduril’s work.

Keene said he joined Anduril after being recruited by a fellow veteran who he served with on the SEAL team, and quickly noticed how many people moved to the defense tech space after successful careers in the government and military. He was promoted to chief of staff at Anduril in June 2022.

“It’s a wonderful opportunity to continue to contribute to the greater defense mission with a company that is taking a completely different approach to defense technology,” he said. “Having the experiences from both the Academy and military service prepared me to understand the needs of the warfighter and what capabilities they need, and how the technology can translate into an advantage.”

He also notes a growing shift toward investment in Silicon Valley as a destination for defense tech. The move is a necessary one, he said, as startups bring a nimbleness and innovation that is critical to staying at the forefront in the cyberwarfare space.
“China is outpacing the U.S. in the AI and autonomous systems,” Keene said. “But high-caliber engineers that normally would be going to top-tier companies like Google, Tesla, Uber and Microsoft are coming to defense companies like Anduril and applying their talents to solve some of the most complex defense problems.”

It’s not just the engineers who are making moves. In 2020, Anduril became the youngest company to win a program of record with the U.S. government since the Korean War. The contract called for Anduril’s autonomous surveillance towers to be deployed. Late last year, Anduril announced it received nearly $1.5 billion in Series E funding—the third largest venture capital raise by a U.S. company in 2022—which will dramatically increase their research and development efforts and further expand their portfolio of capabilities.

“Our software-first approach allows us to move quickly when developing our capabilities,” Keene said. “Then we can go to the government and say, here’s this novel thing we’ve developed that we believe addresses the needs that you have. Let’s continue to work together and refine it, and ultimately deliver it to the warfighter in large enough quantities on a timeline that makes a decisive difference.

“The military and large defense companies are historically good at building exquisite hardware systems like planes, ships and tanks. But the wars of the future are not going to be won with large hardware that takes a ton of time, money and human capital to build. Large quantities of lower-cost unmanned systems and the software that enables them is going to be critical and there simply needs to be more focus on true mission autonomy.”