Roskomnadzor deploys AI to monitor prohibited content

Roskomnadzor uses artificial intelligence algorithms to search for prohibited information on the internet. This was disclosed by the agency’s head, Andrey Lipov, in an interview with «Ъ».

He said that the final decision to block content is taken by people.

“To increase the effectiveness of monitoring, we use neural networks and projects related to artificial intelligence. […] The decision to block is made by authorised personnel, after which the data are transmitted in the format of a registry to telecommunications operators,” Lipov said.

He added that Roskomnadzor’s operators will receive only a portion of the messages. The agency expects that by the end of the year the volume of information that its staff will be able to analyse will increase fourteenfold.

Lipov noted that Roskomnadzor uses both in-house and third-party technological developments.

“Based on the Main Radio Frequency Center (MRFC), we created a scientific council, inviting leading players in Russia’s AI market. Therefore, part of it consists of in-house developments, and part is commissioned,” he said.

Earlier Lipov said that in Russia access to VPN services could be restricted, which would not bring such activity into compliance with the law.

In May 2021, Roskomnadzor threatened to slow traffic to Google and YouTube due to insufficient filtering of materials with information prohibited on Russian territory.

In March, Instagram introduced two AI-based tools, designed to protect young users of the social network from inappropriate messages.

Subscribe to ForkLog’s news on Telegram: ForkLog AI — all the news from the world of AI!