Maxim Ali reviewed for Pravo.ru a neural network being developed by Roskomnadzor to search for banned content

The Roskomnadzor structure has allocated 57.7 million rubles to develop the “Oculus” system of searching for illegal content. Based on neural networks, it will analyse photos, videos and texts on websites, social networks and messengers for prohibited information, including materials with signs of extremism and terrorism, calls for illegal activities, expressions of clear disrespect for society, the state and official symbols of Russia, promotion of non-traditional sexual relations among minors and ways to commit suicide and make drugs.

The creation of “Oculus” should not be seen as a game-changing move on the Internet, Maxim Ali, Partner, head of IP/IT practice at Maxima Legal, said. “From all appearances, “Oculus” will only be an assistant in finding banned content. Ultimately, we are not talking about an artificial intelligence censor. The decision will remain with the government agency, and it will always be verified and signed by a human,” the expert explained to industry portal Pravo.ru.

Maxim also drew attention to the fact that even if a neural network detects illegal content, there is a high probability of false positives. “For example, information on ways to evade responsibility for drug trafficking is considered illegal. How will this information be recognised by the software and distinguished by the neural network from lawsuits, Wikipedia or law degrees?” – he noted.

To read the full article (in Russian) please visit Pravo.ru website >>>