Banks accounted for half of the revenue from AI in Russia

According to the authorities, artificial intelligence has already allowed Russian businesses to gain economic benefits from the implementation of more than 1 trillion rubles. Moreover, according to preliminary estimates, up to half of them come from banks. The financial sector is leading the way in the implementation of AI, which brings quite tangible results, but so far only to the largest players. But, experts warn, the technology also creates risks that may require regulatory intervention.

The economic effect from the introduction of artificial intelligence (AI) technologies in Russia is estimated at more than 1 trillion rubles, Russian Deputy Prime Minister Dmitry Chernyshenko said on March 13. According to his data, a third of Russian companies use AI for work tasks.

“We can confidently say that the financial sector is at the forefront of AI monetization,” says Viktor Dostov, head of the Association of Electronic Money and Transfer Market Participants. “What exactly is the amount of 1 trillion rubles. falls on banks, it’s hard to say, but with a high probability it could be 45–50%,” believes Vadim Tedeev, managing partner of IPM Consulting.

According to the head of the Financial Innovations association, Roman Prokhorov, according to the results of the research, Sber and Yandex can now be called leaders in the Russian Federation in the field of AI.

“The gap with other market participants is large,” he emphasizes. “But in general, the use of AI in the financial sector is one of the leading areas. Accordingly, the share of the banking sector in the effect of introducing AI is close to 50%.”

Sberbank plans to receive 400–450 billion rubles in 2024. economic effect from the introduction of AI technologies versus 350 billion rubles. in 2023. Today at the bank, AI makes 100% of decisions regarding lending to individuals and 40% in lending to corporate clients; by the end of the year, the bank intends to increase this figure to 70%. Sberbank clarified that investments in this area have already paid off, in particular, at the end of 2023, each ruble invested in AI brought about 8 rubles. arrived.

Other largest banks do not disclose such detailed indicators, but confirm the active use of the tool. “AI technologies are involved in all processes of the bank. They are built into anti-fraud solutions, credit scoring, risk assessment, loan repayment, the client’s propensity to buy or fall under the influence of scammers, etc., say VTB. The introduction of AI is part of a large-scale digital transformation project at VTB, economic the effect of which already exceeds 300 billion rubles.”

According to various estimates, the use of AI systems can increase labor productivity in certain areas of banking up to 30 times, says Mr. Tedeev. “AI technologies require banks to make large investments in creating and debugging their information support, but subsequently provide returns by reducing costs,” agrees Managing Director of the NKR rating agency Stanislav Volkov. “As with other investments, the profit from the implementation of AI is higher, the larger its application.” Therefore, small banks, even having developed a successful solution, sometimes cannot make a significant profit due to insufficient scaling, the expert adds, and “they are afraid to share technology with competitors.”

However, market participants also recognize the risks associated with the introduction of AI, which are divided into two groups – reputational, ethical and operational. The first is a challenge for all sectors of the economy, not only for financial institutions. Operational risks are clearly manifested specifically among credit institutions. “This, for example, is the incorrect operation of the models, so we put a lot of effort into monitoring and validation,” explained Alfa Bank. There, as in Sberbank, “in retail, 100% of issuances are already based on AI, in the corporate segment – less than half.” The main task of the bank is to increase this share.

“Banks with less financial resources and focused on the classical business model are less willing to invest in AI, since the maintenance of automated processes requires high costs,” clarifies Vadim Tedeev. “AI also remains a complex system, susceptible to various attacks, that is, there are increased risks of data leaks and disruptions in the operation of automated processes, which can lead, for example, to disruptions in customer service or even losses.” Among the risks, Mr. Prokhorov notes the problem of “depersonalization of responsibility for erroneous AI decisions and the possibility of violating customer rights and competitive conditions.” This, he emphasizes, “requires some regular intervention.”

Ksenia Dementieva, Maxim Builov