Subordinate to Roskomnadzor, the Main Radio Frequency Center (GRCHTs) announced a tender for the creation of the Oculus system based on artificial intelligence (AI), which will detect prohibited content in photos or videos, such as extremism or pornography. The corresponding proposal was published on the website of state purchases, the first to draw attention to the documents was the newspaper “Kommersant”.
Contract price – 15 million rubles. The tender was announced for “the implementation of research work to study the capabilities of artificial intelligence for the automated detection of signs of violations of the legislation of the Russian Federation in images and video materials.”
The system must be able to recognize in a photo or video not only prohibited symbols, but also text. The deadline for fulfilling the contract is December 17, the system is planned to be launched in 2022.
The GRCHTs explained to Kommersant that photo and video materials on the Internet prevail over text and “it is in such content that more and more information prohibited by the laws of Russia appears”.
The head of Roskomnadzor, Andrei Lipov, noted in May that a scientific council was created on the basis of the GRChTs, which included the leading players of the Russian market in the field of artificial intelligence. “To improve the efficiency of monitoring, we use neural networks and projects related to artificial intelligence. Some are our own developments, and some are custom-made. We are trying to make a modular system to make it easier to develop, to use different subsystems for different purposes, ”he said then.
According to Lipov's forecast, by the end of this year, the amount of information that Roskomnadzor employees can analyze for violations will increase 14 times thanks to AI.
In early February, Roskomnadzor launched a mobile application for filing complaints about prohibited content on the Internet. The application made it possible not only to leave a complaint, but also to track the stages of its consideration and check whether access to the specified Internet resources was restricted. The department believes that the application “will increase the effectiveness of protecting Internet users from socially dangerous content.”
