Financial Sector Revamps Security Measures Against AI Hacking Threats

Financial authorities are expanding threat information sharing among financial institutions and pushing for the lifting of network separation regulations to combat hacking using artificial intelligence (AI). Experts advise that while improving the expertise of security personnel and response procedures is crucial, the timing for lifting network separation should be adjusted for sensitive data and services that are difficult to protect.

According to a security system innovation plan announced by the Financial Services Commission on November 5, a system will be strengthened to quickly share signs of attacks detected by financial companies with other institutions. This will include not only threat information identified by the Financial Supervisory Service and the Financial Security Institute but also data on attempted breaches detected by individual financial firms. The goal is to implement this system in the first quarter of next year after a pilot operation next month.

The frequency of vulnerability assessments will also increase for large financial companies, from once a year to twice a year. For smaller firms that lack personnel and budget, the Financial Security Institute will use AI to identify vulnerabilities at external points and provide prioritized actions and response strategies based on risk levels.

The key challenge is keeping pace with the speed of attacks. An information security expert emphasized, "Attacks using AI occur at machine speed, and if humans take time to verify and make decisions, we risk missing the golden time for response." Therefore, it is essential to enhance both the expertise of security personnel and the response procedures.

To bolster AI-based defense capabilities, financial authorities are also considering a complete lifting of network separation regulations. The Financial Services Commission plans to form a task force to discuss this issue. Network separation, which keeps internal work networks separate from external internet networks, has been criticized for hindering the use of generative AI and cloud services.

However, there are recommendations to consider the necessary protection levels for different data and services. Professor Lim Jong-in from Korea University’s Graduate School of Information Security stated, "While the principle should be to lift network separation, we should approach it by delaying the lifting for specific data or internal services that are currently difficult to protect adequately with the existing defense levels."