DoD Gets Serious About AI and Simulation in Wargaming

Pentagon leaders say they’re serious about getting ahead in the artificial intelligence (AI) game, which increasingly could include the “games” involved in the modeling and simulation programs used for training.

Substantial investments in AI by China and Russia have sparked concern that the United States is falling behind in technology advancements that could, as Russian President Vladimir Putin has claimed, determine who rules the world. The Department of Defense (DoD) is putting billions of dollars into AI and supporting technologies, like cloud computing and big data. While some observers question whether it’s enough, others say DoD should not neglect investment in the developing modeling and simulation technologies.

At a recent National Training and Simulation Association conference in Orlando, Fla, Tony Cerri, director of data science, models, and simulations at U.S. Army Training and Doctrine Command G2, encouraged industry to incorporate AI into the simulations that are foundational to modern military training. Cerri said such an approach would set the U.S. military apart from what other countries, such as China and Russia, are doing, and give DoD what amounts to “an unbeatable advantage,” National Defense reported.

The military has always stressed the importance of realistic training, from early flight simulators dating to the 1920s to the Army’s Live Synthetic program, which combines virtual worlds with live-fire exercises. Ten years ago, the Army invested $50 million in its PC-based multiplayer Games for Training application, and has followed up with additional contracts to develop more realistic avatars, sharper displays, and better multicast networking. These investments and other simulations have proved useful, but the changing landscape of real-world conflicts requires more. Proponents say the next generation of training scenarios will need to make greater use of AI and big data because those technologies are going to be employed in any real conflict.

Alexander Kott, chief scientist at the Army Research Laboratory, points out in a recent post explaining AI programs that AI is essential to creating realistic immersive simulations that allow the Army to follow its “train-as-you fight” principle. These exercises include creating realistic physical and sociocultural interactions with simulated agents, while operating in a complex setting that mirrors the real world–a world in which AI agents also will be involved. After all, human-machine interaction won’t be limited to training scenarios, Kott writes. “The actions of human actors teaming with robots and other intelligent agents will be pervasive in the complex operational environments of the future.”

Incorporating machine learning and other cognitive capabilities into training scenarios can help forces become effective in the multifaceted environment of hybrid warfare. Russia has been at the vanguard in these nonlinear tactics, employing cyber and economic attacks, propaganda campaigns, covert troops it claims are local defense forces, and other strategies to destabilize a region or government. Training programs need advanced analytics to keep up, the thinking goes. The European Defense Agency, for example, has launched a one-year study into applying AI and big data in training and simulations, using methods such as data farming and war gaming specifically to analyze how to deal with “very complex scenarios like hybrid warfare.”

Tech Sector’s Not Playing Games

Many of the leading companies in AI and simulation are dead serious about this DoD simulation business.

“AI will touch every aspect of simulation and training, from creating the synthetic environment to leading the opposition forces and creating challenging scenarios that push the war fighter to learn new skills,” said Anthony Robbins, vice president of public sector at NVIDIA. “The potential to use deep learning to analyze after-action reviews and glean insights from huge volumes of data created across the training career of a soldier or a combat brigade will improve training and readiness in ways that just weren’t possible before.”

“As DoD incorporates AI into its operational capabilities, its critically important training should not be an afterthought,” said Charles Waters, account lead for the Army Logistics and Cloud Services at IBM. “AI-supported, immersive training increases the speed at which trainees and soldiers learn and master evolving warfighting techniques and technologies. As geographies, actors, and agendas change among conflicts, both potential and real, AI can modify training environments without massive net-new training content creation. And beyond enhancing our response to warfare, AI and supporting analytics can assist in the development of new tactics to deploy.”

While industry is bullish about these new capabilities, it provides a cautionary note and stressed the need to think for the future. “Legacy storage is a bottleneck when it comes to turning big data into big advantage through real time intelligence,” said Gary Newgaard, vice president public sector at Pure Storage. “Big data will play a key role in developing trainings and simulations for the military and a next-generation data platform will be essential to supporting these efforts. Systems for the future cannot be designed on the technology foundation of the past.”

Move over Xbox Call of Duty–here come the AI Marines.

Recent