The Air Force is using machine learning through a pilot program to reduce unscheduled maintenance by as much as a third, while the Army Research Laboratory cautions against falling prey to “snake oil” pitches on artificial intelligence.

After ingesting about seven years’ worth of data from a single airframe platform,  the Defense Innovation Unit and the Air Force found about a 30 percent decrease in unscheduled maintenance for its aircraft.

Mike Madsen, the Defense Innovation Unit’s director of strategic engagement, said those results could save time, money and lives.

“In the aircraft maintenance world they collect huge amounts of data, but they’re just now learning how to how to manipulate this for things like predictive maintenance,” Mike Madsen, the Defense Innovation Unit’s director of strategic engagement, said Wednesday at a GovernmentCIO forum on AI and big data in Arlington, Virginia.

The military services have already taken several steps in using AI to determine what equipment needs repair before it’s broken.

This summer, the Army, working alongside DIU, reached a prototype award with Chicago-based contractor Uptake for its Bradley Fighting Vehicle.

Under this award,  the Army will use Uptake’s Asset Performance Management application to predict component failures and decrease the frequency of unscheduled maintenance.

In a 2016 pilot program, the Army Materiel Command’s Logistics Support Activity partnered with IBM to use its AI platform Watson to keep track of the maintenance needs of 350 Stryker vehicles.

Madsen said the services continue to look for ways to leverage AI tools.

“The Marine Corps just called me the other day,” Madsen said. “They want to look at applied predictive maintenance to some of their vehicles. There’s an opportunity to scale across platforms as well as the services.”

NGA delivering on AI ‘from the government out’

While DIU sees lowering barriers into the defense marketplace as part of its mission, the National Geospatial-Intelligence Agency looks to develop AI tools internally, and build a workforce with the skills to take emerging technology to the next level.

“We wanted to revitalize the skill set and the ownership, from the government perspective, of having federal employees that can code,” as well as develop software-as-a-service applications, said Todd Myers, NGA’s automation lead.

But NGA also wants employees who understand the history of what they’re doing — from the pioneering days of Alan Turing, the father of computer science, to the present day.

“There’s context there, and it’s been completely lost. And you don’t get that by buying AI and [machine learning] from a company — you get that from developers in government taking ownership and developing against context for mission relevancy,” Myers said.

NGA has also spent considerable time over the past two years building inroads with innovation hubs like Silicon Valley, and looking for ways to work with that code into the government space.

“We’re bringing the ownership in on my team, to deliver from the government out, which is converse to a lot of the way things are done,” Myers said.

Don’t fall for the AI ‘snake oil’ pitch

For all the hype around AI, Alexander Kott, the chief scientist of the Army Research Laboratory, said agencies should set realistic standards for AI, and not fall prey to “irrational exuberance” over what AI can do for them.

“Let’s not be the part of the snake-oil seller crowd for AI. Let’s be realistic about AI — let’s focus on what it can do for the specific tasks speed of specific missions of your organization,” Kott said. “If it’s appropriate, it’s appropriate. If it’s not appropriate, don’t sell it.”

Although AI often gets mentioned in the same breath as emerging technology tools like blockchain, it’s been discussed for decades. Kott began working with AI in the 1980s, adding that AI in government isn’t a fad, but it’s also not the right solution to every agency’s data and technologies woes.

“If it is already being attempted to be automated, you probably need to ask yourself: Why did it fail before? AI is not a magic pixie dust that you can sprinkle on your problem and it suddenly will get solved,” Kott said.

Like the rest of the government, the Army Research Lab also competes for in-demand talent —particularly individuals who’ve recently earned their PhDs in science, technology, engineering and math.

“We do experience a lot of market competition, and in some hot fields, we really do see almost an inability to attract some of the important demographics,” Kott said. “At the same time, if you focus on the uniqueness of your mission, you usually find people who will love that particular mission.”

While the federal government struggles to compete with private-sector salaries, Kott said the lab’s mission continues to attract a steady stream of scientists.

“When people come to the Army Research Laboratory, we talk about the opportunity of doing amazing science — science that can influence a lot of lives which can save soldier lives, science that is going to make or break the future greatness of this country.”

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.