The key facts
- Maxim Oreshkin assessed the lag of Russian AI models behind American and Chinese ones at 6–12 months
- According to him, domestic neural networks better account for the linguistic and cultural specifics of the countries where they are used
- Yandex and Sber fine-tune models together with partners from friendly countries
Assessment at the media forum
Deputy head of the presidential administration of the Russian Federation Maxim Oreshkin, at a meeting with participants of the Russia — Africa media forum said that Russian artificial intelligence models lag behind American and Chinese ones by roughly 6–12 months. At the same time, according to him, domestic neural networks better account for the linguistic, cultural and historical specifics of the countries where they are used and are able to prioritize local content in local languages.
Adapting models to local markets
Oreshkin linked the advantage in understanding local specifics to a more precise knowledge of a particular country's characteristics. Yandex and Sber, he said, apply this adaptation experience in cooperation with friendly countries: together with partners, the companies fine-tune models and create solutions designed to work more accurately with local conditions.
AI in Russian courts
Life.ru also recalls a previously published report according to which Russian courts began testing artificial intelligence in more than 20 regions of the country. Neural networks are planned to be tasked with analyzing documents, searching for errors and keeping records of hearings; as noted, they will not make decisions on cases.




