New War Revolution! Ukraine trains military AI with 2 million hours of video, can it win against Russia?

Large-scale image data 'feeding' AI makes precise decision-making possible

In the fierce battlefields of the Ukraine-Russia conflict in the past two years, thousands of drones with cameras have been flying day and night, recording over 2 million hours of battlefield footage, equivalent to about 228 years of video length. Nowadays, Ukraine is treating these massive aerial data as an "AI feeding source" to train military artificial intelligence in image recognition and automated decision-making. This initiative is led by the non-profit organization OCHI, aiming to combine real-time drone imagery with AI analysis on the backend to assist commanders in obtaining a more accurate understanding of the situation or even allowing the system to detect and attack targets on its own.

OCHI founder Oleksandr Dmitriev emphasized that these aerial videos contain unparalleled value: 'Giving AI 2 million hours of battlefield records can allow it to quickly self-learn in many aspects such as tactics, target identification, and weapon system optimization, becoming an unimaginable battlefield intelligence engine.' Dmitriev said that after extracting common patterns through AI simulation algorithms, the system can determine the best shooting angle or route and provide dynamic suggestions to the commander.

Source: Reuters Ukrainian soldiers from the 24th Independent Mechanized Brigade 'La Roga' install anti-tank mines on a 'Vampire' combat drone.

From commanding support to autonomous attack, the AI-ization of drones is becoming the norm

In fact, drones have become an indispensable part of modern warfare. In the past few years, Ukraine has gradually deployed a large number of commercial or modified models to deal with the Russian army. With the advancement of AI technology, some drones can use deep learning to automatically distinguish terrain, avoid obstacles, and even judge attack targets on their own. The industry points out that if all the videos in the massive database are included in AI training, allowing machines to infer decision rules from hundreds of thousands or even millions of combat scenarios, drones may be able to complete complex tasks without human intervention in the future, further shortening response time.

According to US think tanks and media reports, not only Ukraine, but also the Russian military has invested in a variety of AI applications, including the automatic target recognition function of the Lancet series UAV. At the same time, the Ukrainian army is also developing the concept of "AI group flight": through a central computer system to control dozens or even hundreds of drones at the same time, a series of "cloud and cloud" joint combat formations. In the future, combined attacks that break out instantly on the battlefield will be even more unpredictable and defenseless.

The debate between data volume and quality: combating different operational environments

Although there are 2 million hours of video resources, AI still relies on data quality to achieve high-precision detection and judgment. Kateryna Bondar, a researcher at Wadhwani AI Center, said that AI models need to accurately recognize terrain, colors, and shapes, which means that image clarity and annotation quality are crucial. 'Even with a large quantity, the system cannot accurately identify if the image is just a blurry black and gray screen,' she also pointed out that the United States and other countries are interested in this, but they are more interested in collecting images that can be used to counter potential Asia-Pacific war environments (such as islands, seas, and tall buildings in cities), which will be beneficial for future applications in scenarios involving China.

On the other hand, the accumulated huge amount of videos also poses a considerable challenge in terms of security protection: if they are stolen or destroyed by the enemy, AI training materials will be greatly reduced. Dmitriev emphasized that OCHI and another system, Avengers, of the Ukrainian Ministry of Defense are dedicated to encryption in different storage and transmission processes to safeguard valuable information.

Image source: AFP The artillery unit of the 128th Mountain Assault Brigade of the Ukrainian ground forces is training on how to use drones.

The next step in the militarization of AI: the dialectic of efficiency and morality

The combination of a large number of drone images and AI algorithms may bring revolutionary battlefield efficiency: from real-time calibration of enemy equipment to joint strikes, it can be completed without the traditional "observation-confirmation-command-execution" layers. However, the position of the "human" in the chain of command and the potential risk of miscalculation are also being discussed due to the increased degree of automation. If decision-making power is greatly transferred to machines, how can accidental bombing of civilian facilities or disputes over the ethics of war be avoided? There is still no clear international consensus on this.

At the strategic level, the massive drone data collected by Ukraine is undoubtedly a 'dream training ground' for AI training, and many military industrial enterprises in Europe and America also want to take the lead in this technology. When all forces gather, AI will surely develop rapidly in the military industry and influence the future military landscape evolution.

As Dmitriev said, "Machines can transform thousands of experiences into mathematical calculations, and as long as we have enough visual data, we can teach them how to fight."

Overall, Ukraine's active collection of drone battlefield videos marks the beginning of a deeper integration of AI and the military. If the technology matures, drone swarms equipped with AI engines will become the key to determining the outcome of battles. In the future, similar patterns are likely to be replicated and accelerated in conflicts in Europe or other regions, challenging the international community's thinking on norms of warfare, privacy data, and ethical risks.

[Disclaimer] The market is risky, and investment needs to be cautious. This article does not constitute investment advice, and users should consider whether any opinions, viewpoints, or conclusions of this article are suitable for their specific situation. Invest at your own risk.

The New War Revolution! Ukraine uses 2 million hours of video to train military AI, can it win against Russia? This article was first published in 'Cryptocity'.