Russia’s Ministry of Defense has developed a new platform to train military artificial intelligence using footage captured during drone combat operations. The service gives developers access to battlefield video records that can be used to build and improve AI systems, the ministry announced.

At the center of the project is a dedicated video repository containing prepared recordings from FPV drones and other unmanned aircraft, including both multirotor and fixed-wing models. Developers can use this material to train neural networks designed for military applications.

To ensure that participating organizations follow a consistent approach, ministry specialists have introduced a unified data classification system. It provides a common framework for organizing information and training AI models according to shared standards, TASS reported.

The platform also allows developers to submit the results of their work back into the system. Uploaded data undergoes checks for completeness and compliance with the established classification format. The Defense Ministry can then use the collected information to develop and train its own military neural networks.

More than 40 organizations have already joined the project. Russian Deputy Defense Minister Dmitry Shcherbinin, who oversees its implementation, said the participating companies and institutions are training neural networks, including models intended for integration into their own combat drone systems.

Information security is another component of the platform. According to the ministry, the service uses certified security solutions, modern cryptographic technologies and authentication mechanisms to protect its data.

The Russian Defense Ministry expects the system to strengthen the resilience of the country’s Armed Forces and improve their ability to adapt to emerging challenges in artificial intelligence.