Roskomnadzor launched the Oculus system - it automates the search for information that causes complaints from the department. This was reported by Vedomosti with reference to a representative of the Main Radio Frequency Center (GRFC), subordinate to the RKN and which was the customer of Oculus.
As a representative of the GRChTS told the publication, the system is already fully operational. In December 2022, it was tested, and at the beginning of 2023, it was linked with other RKN monitoring systems.
Plans to create such a system became known in September 2021. developing Oculus for 57.7 million rubles began In August 2022 , Extraction RDC . According to the tender, the system analyzes images and videos, URLs, chats and other materials in order to find content that contains signs of extremism and terrorism, calls for uncoordinated rallies, expressions of disrespect for the state and its official symbols, information on methods of committing suicide, as well as propaganda of smoking, drug production, “non-traditional values” and LGBT people.
It is expected that Oculus will significantly simplify monitoring for the RKN - previously, GRChTs employees searched for information for the department manually. “On average, operators processed 106 images and 101 videos per day. Oculus will analyze more than 200,000 images per day (about three seconds per image). The system will automate and significantly speed up the monitoring of visual content,” added Vedomosti’s source.
By 2025, they plan to improve the system by adding new classes of violations and other functions that will allow identifying prohibited materials “in several frames of video fragments, in complex handwritten texts and drawn content.”
In November, the Belarusian group “CyberPartisans” gained access to internal documents and letters of the Main Radio Frequency Center (GRFC). “Mediazona” analyzed the received archives and told how RKN, with the help of the Moscow Institute of Physics and Technology and the Brand Analytics company, monitors content on the Internet. You can read about what else has become known from the GRChTs archive here .