T-Bank launched a cyberattack on its own infrastructure using artificial intelligence
Russia's T-Bank has become the first organization to use artificial intelligence technologies to test the resilience of its infrastructure against cyberattacks. For this purpose, a special system called Nulla was developed. It consists of offensive agents capable of automatically identifying vulnerabilities without human intervention. This was reported by Ixbt.com reports.
During the testing process, over 1,300 services and platforms, including T-Business, T-Travel, and T-Auto, were audited. Thanks to the efficiency of the new system, the time required to test a single service was reduced to 45 minutes. For comparison, a traditional manual audit process usually took 2–3 days.
Unlike classic security tools, the Nulla system automatically adapts its attack strategy to the specific features of the defense system and builds complex vulnerability chains. This solution is expected to provide an economic benefit of 100 million rubles by preventing potential cyber incidents.
T-Bank representatives noted that the system has high scalability and can analyze dozens of services simultaneously without losing quality. This project, which began in April 2025, is already being actively used to protect the bank's ecosystem.
Additionally, in June 2026, T-Bank will launch new cybersecurity tests. A prize pool of up to 12 million rubles has been allocated for participants who execute critical scenarios.
