The National Bank of Kazakhstan plans to shift from investigating crimes after they occur to preventing fraud before money is transferred. To achieve this, the regulator will combine data from banks, telecom operators, marketplaces, and other market participants, with artificial intelligence identifying suspicious activities. This was reported by Qazaqyia.kz citing Kursiv Media.

This was announced by Deputy Chairman of the National Bank Binur Zhalenov.

"We must admit and speak directly – the problem of fraud is still very relevant today. We, like government agencies in general, have great potential to work and improve our efficiency," he said.

According to Zhalenov, the fight against fraud is currently mostly reactive. Measures are typically taken after money has been transferred, the victim has contacted the bank or police, and then an investigation begins to recover stolen funds.

Now the National Bank aims to build a proactive system that can identify fraudulent schemes before a suspicious transaction is completed.

To this end, starting in 2026, the regulator is merging data from telecom operators, financial organizations, marketplaces, cryptocurrency infrastructure, and other sources.

"Fraud is usually generated through calls, correspondence, i.e., through telecom operator data. Therefore, starting this year, we are merging data from telecom operators, financial operators, marketplaces, crypto assets, etc., to use artificial intelligence to proactively identify potential fraudsters and, most importantly, potential victims based on behavioral patterns," Zhalenov explained.

According to him, the system will analyze user behavior. If a person is under the influence of social engineering, their actions often become atypical.

"When a person is under the influence of social engineering, they begin to act abnormally – opening many banking apps, making calls, and so on. The same applies to fraudsters. We are now studying artificial intelligence so that it can automatically detect such patterns," noted the Deputy Chairman of the National Bank.

When suspicious activity is detected, the system can send signals to banks or telecom operators to take measures to prevent fraud, including blocking transactions or other protective mechanisms.

The National Bank's presentation also noted that the new system plans to use a pair of AI agents. One, the "Analyst," will study each fraud incident and generate recommendations for investigators. The second AI agent, the "Observer," will monitor the internet and the Darknet to identify new fraud schemes and typologies.

The Darknet is a hidden segment of the internet that is not indexed by regular search engines like Google. It requires special software (e.g., the Tor browser) and ensures complete anonymity. The network is used for both illegal trade and protecting free speech.

The National Bank expects that using artificial intelligence will allow a shift from investigating crimes already committed to their timely prevention.

According to Zhalenov, a pilot project has already been launched, and its first results are planned to be presented by the end of 2026.