Online fraud cases in Kazakhstan fall by 40%

Over the past three years, the number of online fraud cases in Kazakhstan has fallen by 40%, while the detection rate has increased by 29.9%.

Since the beginning of 2026, 14,414 online fraud cases have been registered, 13% fewer than during the same period last year. Victims received ₸5.9 billion in compensation.

The National Anti-Fraud Centre has become one of the main tools for combating fraud. It has operated since July 2024 and connects more than 200 participants, including banks, payment organisations, government and law enforcement agencies, and telecommunications operators.

Almost 180,000 suspicious incidents were registered in the system over two years. By sharing data, banks prevented the theft of more than ₸117 billion and returned over ₸578.5 million to victims.

The Anti-Fraud Centre allows participants to quickly exchange information about suspicious transactions and block funds before fraudsters can withdraw them.

New requirements for protecting personal data and digital infrastructure are now in effect. Organisations must notify citizens and government agencies of data breaches.

Biometric authentication is required to access databases containing medical, banking, and other confidential information. Owners of digital infrastructure must also conduct annual cybersecurity audits.

Since 2024, authorities have conducted 648 unscheduled inspections, including 480 related to information security and 168 to personal data protection.

Citizens can use the Stop Credit service on eGov.kz and in the eGov Mobile app to voluntarily block loans from being issued in their name. By 1 June 2026, the service had been used 6.3 million times.

Other available tools include a service for managing access to personal data and the NomadGuard platform for checking potential data breaches and phishing links.

Experts recommend keeping SMS codes, passwords, and bank card details private, checking links before opening them, and using two-factor authentication.
ChatGPT может допускать ошибки. Данные рабочей области не используются для обучения моделей.