AI is already learning from Russia’s war in Ukraine, DOD says
Pentagon AI tools are generating battlefield intelligence for Ukraine. An unusable Russian tank on the Kyiv-Zhytomyr highway after Ukrainian soldiers recaptured the region in early April 2022. Metin Aktas/Anadolu Agency via Getty Images
Today’s battlefield data is helping smart machines model the wars of the future.
Less has been said about the use of artificial intelligence in the Ukraine war than, say, anti-tank missiles, but the Pentagon is quietly using AI and machine-learning tools to analyze vast amounts of data, generate useful battlefield intelligence, and learn about Russian tactics and strategy, a senior Defense Department official said on Thursday.
“What you're not seeing,” said Maynard Holiday, director of defense research and engineering for modernization, is “our exquisite intelligence capabilities that are able to oversee the battlefield,” including gathering and archiving signals intelligence.
“We'll definitely be doing an after-action analysis on everything we've seen with respect to Russian tactics,” Holiday said at Defense One’s Genius Machine’s AI summit on Thursday, and all of it will go into a database that “we can train and then war game on.”
Just how much battlefield intelligence the U.S. is passing to Ukraine is a matter of some conjecture. The United States is not operating drones in Ukraine, but commercial satellite companies have made large volumes of pictures and images available to the public.
Also on Thursday, Gregory Allen, who leads the AI Governance Project and is a senior fellow at the strategic technologies program at the Center for Strategic and International Studies, pointed out that the military’s AI tools for finding and tracking specific objects in footage collected from drones had advanced considerably over the last several years. And the military has begun doing the same with satellite photos.
Allen said military AI has come a long way since 2017, when the public learned of Project Maven, the military’s object recognition program.,
“Artificial intelligence [and] machine learning has become an increasingly capable and increasingly widespread factor in United States intelligence, surveillance and reconnaissance operations,” he said. It’s been quite useful for tracking what’s going on in Ukraine. “The United States Department of Defense and our allies are taking advantage of what's been built over the past five years.”
In other words, applying advanced AI tools to publicly available imagery is producing critical information to help the Ukrainian military thwart Russian attacks.
This war’s data will help the military better model and anticipate how an advanced adversary will behave in the real world, particularly Russia and China. It’s something that military leaders have said needs to start happening now.
Holiday said he was working on that with the Naval Information Warfare Center, Pacific, in San Diego, California. The Center’s Battlespace Exploitation of Mixed Reality Lab is seeking to understand how rapidly advancing technology will shape adversary behavior. Most wargames focus on current technology and behavior.
The United States needs to be able to model changes over time, to look at what the Pentagon and neer-peer adversaries will have in 2025 or 2030, he said.
"We'll be able to visualize all of that, you know, for leadership and then also for multifactor analysis between our different technology areas.”
Russia’s lackluster performance in Ukraine and the crippling sanctions that Western states have placed on the country suggest that Russian AI development will slow down—but not stop.
“Even if Russian economy and even if the technical ecosystem continued to be battered by the sanctions, this will not stop the MLD thinking through what these concepts may eventually mean once the war ends one way or another,” Samuel Bendett, an adjunct senior fellow at the Center for a New American Security and an Adviser at the CNA Corporation, said on Thursday.
NEXT STORY: DIU sets up shop in Chicago