JUST IN: Militaries Behind in Applying AI to Training, Simulations

iStock illustration

LONDON — Despite artificial intelligence being a hot topic among military technologists for years, the training community is only in the beginning stages of exploring how to apply it to their high-tech simulators and modeling software, officials said April 26 at the IT2EC trade show.

“AI in our professional lives is still attached to the words ‘innovative’ and ‘new,’ and yet, we use our phones, we use this technology all the time,” said Air Vice Marshal Ian Gale, director-general of the Defence Academy and Joint Force Development at the U.K. Strategic Command.

“There is this stark difference between our home lives and and what we can achieve and our military lives,” he said in a keynote speech at Europe’s largest training and simulation conference.

The U.K. Ministry of Defence is currently developing a “future operating concept” and AI will play a role, he said. Some of the tasks it’s currently performing can seem “boring,” he added, such as sorting through bulk data and translating languages.

But language translations are proving useful, he said, noting that AI is helping translate the thinking of rival nations.

“There are two nations that are spending the most most money on AI and it’s likely they will come to the best technological solutions. We don’t necessarily need them. We need to be good enough,” he said of the United Kingdom.

Maj. Gen. Richard Oppelear, commander of Netherlands Defense Academy, said “we presume that all our youngsters and their generation are experts in AI. I don’t think they are,” he said during a panel discussion. The Netherlands doesn’t have a program or class on artificial intelligence.

“You don’t have to become a data and analytics specialist, but you have to cope with the environment and the tools you have available,” he said.

AI is only as good as the quality of data that feeds it, the officials noted. Data must be must be standardized before it can be input into training systems to make them more robust, they said.

U.S. Army Col. Heath McCormick, commander of the Joint Multinational Simulation Center in Grafenwoehr, Germany, said AI can be particularly useful in wargaming. It can rapidly provide feedback on actions taken.

AI can speed up the decision making process during wargames. “There’s a huge opportunity there,” he added.

“Yet,” he said “it’s just a tool we are all going to use and I don’t want to be a slave to it.”

The biggest impediment to using AI in training and simulations is a lack of data, he said. “The data is out there, but we can’t harness it,” McCormick said. “We’ve got to have clean data and data that is relevant in the right formats.”

Gale said the world is full of AI. “It’s how we get the correct AI harnessed to the best of its ability to train the the warfighter in training and education and in real operations.”

As far as real operations, the U.K. Ministry of Defence was able to make model of a Russian invasion of an “Eastern European country” to see how heavy armor vehicles get bogged down in mud based on weather and terrain data. It also did something similar with fuel and logistics to see how that would affect Russian trucks, he said.

“We might be a long way behind [on AI] but we’re sort of ahead of the competition,” Gale added.

Karen Saunders, the U.S. Army’s program executive officer for Simulations, Training and Instrumentation, said “We need to understand the data that we need for our simulations to consume.” The PEO also needs to understand its military customers’ needs. “You need to understand what their data requirements are,” she said.

And once it has that data, the PEO needs to know what to do with it. How to turn it into information to develop training tools, she added.

“I think that’s where we can … look to industry where such subject matter experts can talk to us about different techniques that [they may] have for processing data into actionable information,” she said.

Topics: Training and Simulation, Robotics and Autonomous Systems