As the future of warfare moves toward artificial intelligence, Ukraine is sitting on a valuable resource: millions of hours of footage from drones that can be used to train AI models to make decisions on the battlefield. AI was used by both sides on the battlefield during Russia’s invasion of Ukraine to identify targets, capturing images much faster than a human.
Oleksandr Dmitriev, founder of OCHI, a Ukrainian non-profit digital program that centralizes and analyzes video footage from more than 15,000 front-line drone operators, told Reuters that his program has collected 2 million hours, or -228, of battlefield video from drones from. 2022.
That will provide valuable data for the AI ​​to learn from.
“This is AI food: If you want to teach AI, you give it 2 million hours (of video), it will become supernatural.”
According to Dmitriev, the footage can be used to train AI models in combat tactics, target recognition and test the effectiveness of weapon systems.
“Essentially experience can be turned into statistics,” he said, adding that an AI system can learn the methods and angles at which weapons are most effective.
The system was developed in 2022 to give military commanders an overview of their battlefield locations by showing them drone images from all nearby teams on a single screen. After the system was rolled out, the running team realized that the video sent back by the drones could be useful as a record of the war – so they started to save it.
On average, Dmitriev said five or six terabytes of new data are being added every day since the conflict.
Image quality
Dmitriev said he has been talking to representatives of some of Ukraine’s foreign partners who have expressed interest in his OCHI program, but declined to provide details.
Samuel Bendett, a senior partner at the US-based New American Security Institute, said that large amounts of data can be very important in teaching AI programs to identify what they are actually seeing, and what actions to take.
“People can do this intuitively, but machines can’t, and they have to be trained on what is a road or not, or an environmental obstacle, or an ambush,” he said.
Kateryna Bondar, a fellow at the Wadhwani AI Center at the Center for Strategic and International Studies, said the size of the data set and the quality of the image are important, as AI models learn to recognize targets based on shapes and colors.
Bondar said the dataset is important in the context of training to fight Russia. However, he said U.S. officials and drone makers are choosing the dataset that trains AI systems to operate in the Pacific against a potential Chinese adversary.
“(They want) plans that are ready and able to fight China because that’s the priority for the US right now, rather than getting a lot of paintings of Ukrainian fields and forests.”
Ukraine also has another program, called Avengers, developed by the defense ministry, which intercepts and collects video from drones and CCTV. The department declined to provide information about the plan. However, it has previously stated that the Avengers detect 12,000 pieces of Russian equipment per week using AI detection tools.
Thousands of drones already use AI systems to fly to targets without human supervision, and Ukraine is using AI technology to help destroy its territory. Ukrainian companies are building a drone network, where a computer program will be able to execute commands for a connected cloud of dozens of drones.
Russia has also touted its use of battlefield AI, particularly target recognition on its Lancet strike drones, which have proved a threat to Ukrainian armored vehicles.
Why should you buy our Subscription?
You want to be the smartest in the room.
You want access to award-winning journalism.
You don’t want to be misled and misinformed.
Choose your subscription package