Russian Central Bank Proposes AI Use to Protect Consumers from Hidden Financial Product Terms

© ЦБ РФ

The Bank of Russia is considering the introduction of unified standards for the use of artificial intelligence (AI) in the sale of financial products, reports the journal "Bankovskoe Obozrenie" (Banking Review). According to Deputy Governor Mikhail Mamuta, modern AI systems can enhance transparency in client interactions, but only if they adhere to principles of neutrality, ethics, and the absence of hidden influence on consumer choices.

Mamuta noted that over the past few years, the financial market has evolved from direct miss-selling and service imposition to more sophisticated mechanisms of client influence. Previously, consumers might have been sold investment products disguised as bank deposits or forced to purchase additional insurance when taking out a loan. Now, companies increasingly employ so-called "native navigation" algorithms, subtly guiding individuals toward specific decisions.

While the regulator does not oppose the idea of client support, the problem arises when consumers are shown only the advantages of a product, omitting its limitations, risks, and additional terms.

One example of such practices, according to Mamuta, is promotional offers for bank deposits with enhanced returns. As citizen complaints revealed, the attractive interest rate was often accessible only if certain additional conditions were met, such as maintaining a minimum account balance, opening a brokerage account, or actively using a bank card. Consequently, the actual terms differed significantly from initial client expectations.

To address this issue, the Bank of Russia, in collaboration with market participants, is testing an AI assistant based on large language models (LLMs). The service is expected to automatically analyze contracts and highlight key product conditions for clients, including potential restrictions, fees, and risks. According to the regulator, the pilot project has received positive feedback from both users and banks interested in long-term client relationships.

In the future, the use of such digital assistants could become an industry standard. Mamuta believes this would reduce the advantages of unscrupulous market players and establish uniform rules for honest consumer information.

Particular attention, he emphasized, should be given to the independence of these systems. It is crucial, according to the regulator, that AI does not become a tool for hidden sales but provides objective information regardless of the interests of a specific financial institution.

Commenting on the development of AI technologies, Mamuta noted that user trust in language models continues to grow. In some cases, consumers are already more willing to accept AI recommendations as more objective than advice from bank employees, believing that algorithms have no personal interest in selling a particular product.

However, Mamuta also pointed out that this creates new risks. Modern language models are increasingly adapting their communication style to users, aiming to elicit sympathy and trust. This "empathy" could lead to excessive reliance on AI recommendations and a reduced critical assessment of the information received.

Another concern raised by Mamuta is the possibility of creating AI models trained in the interests of specific companies. While they may formally compare financial products objectively, they could subtly favor solutions developed by their creators.

Therefore, the state must establish unified requirements for AI operation in the financial sector, Mamuta stressed. These should include principles of neutrality, transparency, mandatory warnings, and disclosure of how AI generates recommendations.

Ultimately, the Bank of Russia believes that the final decision on choosing a financial product should always remain with the individual, with AI serving only as a tool for objective navigation, not hidden influence on client behavior.