The Moscow Exchange has announced the launch of a multi-stage software code verification system based on artificial intelligence. This is a landmark event for the entire financial infrastructure of Russia, as the platform is one of the first among regulators and market operators to begin systematically using AI agents to strengthen cybersecurity.
The architecture of the new solution assumes that a chain of specialized AI agents connects to the analysis already at the zero stage, preceding the work of experts. The system takes on routine operations: scanning software components, cross-referencing with open vulnerability databases, and forming preliminary conclusions. Information security specialists receive a structured report, on the basis of which they conduct in-depth verification and make final decisions.
How the system works
The development was carried out by the exchange's information security team together with experts in technical information protection and practical vulnerability discovery. Within the service, tasks are distributed among several agents:
Analysis — scanning software components;
Cross-referencing — comparing data with open vulnerability databases;
Prioritization — forming conclusions and selecting priorities for manual verification.
This approach allows experts to use the reports as a convenient tool for deep auditing, focusing on the most critical risk areas rather than spreading efforts across routine tasks.
What this gives the exchange
The main effect is high speed and regularity of control. The implemented approach accelerates checks and reduces time spent on standard tasks. The algorithms also contribute to more precise tuning of analysis for the exchange's specific systems. At the same time, platform representatives emphasize: there are no plans to replace people with machines.
"Artificial intelligence does not replace information security experts, but it allows increasing the speed and quality of preliminary analysis; such solutions help process large volumes of data faster, identify potentially significant risk areas, and focus specialists' attention on the most important tasks," noted Sergei Demidov, Director of the Operational Risks, Information Security, and Business Continuity Department at the Moscow Exchange.
The new service continues the platform's strategy of integrating neural networks into control frameworks. Earlier, in July, the exchange already announced the launch of a system for detecting trades with signs of market manipulation and insider trading.
My view: This is a correct and timely step. Financial platforms are the primary target for cyberattacks, and automating primary code analysis with AI is not just a trend but a necessity. However, the key point here is the preservation of human control. Fully trusting neural networks with final decisions in the security domain is still premature, and the Moscow Exchange clearly understands this.