VTB and Central Bank management advocated for lenient legislation for AI models.
5/22/2026, 02:00 PM • Богдан Семичев

Andrey Kostin, CEO of state-owned VTB Bank, openly opposed the introduction of strict restrictions on the use of foreign artificial intelligence models in Russia. Speaking at a specialized banking conference on Friday, the top manager emphasized the need to develop the most liberal legislation in this area with a minimum of restrictions. The position of the head of one of the country's largest credit institutions was fully supported by Central Bank Governor Elvira Nabiullina, who expressed solidarity with the policy of lenient regulation of the industry.
The head of the Central Bank specifically drew the attention of the professional community to the serious risks of overregulation of the technology sector, which requires only balanced and deliberate decisions. Previously, Vadim Kulik, Deputy Chairman of the VTB Management Board, reported that the financial institution actively relies on the core of the Chinese Qwen neural network, as well as on developments from the Russian company Yandex. At the same time, according to experts at the state bank, existing Russian AI platforms from Yandex and Sberbank, at the current stage of technological development, are still inferior in effectiveness to advanced Chinese counterparts such as Qwen or DeepSeek. Responding to legitimate questions from conference participants about the potential threats of unauthorized leaks of sensitive commercial information to China, Andrey Kostin countered with a wry remark that the bank does not possess state secrets and is not part of the Federal Security Service.
Such active discussion of the topic among senior financial sector officials is prompted by recent initiatives from relevant agencies. The Ministry of Digital Development has already prepared a specialized bill regulating artificial intelligence, which envisages the introduction of a number of significant barriers. Specifically, the document proposes direct restrictions on the operation of foreign LLM models within the country, as well as the creation of a separate registry of so-called "trusted" platforms, mandatory for use by government agencies and strategically important sectors of the economy.
