In the future, financial companies will conduct security tests using AI for security purposes Even i..

In the future, financial companies will conduct security tests using AI for security purposes Even if a computer failure occurs, sanctions by financial authorities, such as fines, will be exempted. This is to support the financial sector to actively respond to the security threats of frontier AI such as Mitos.

On the 2nd, the Financial Services Commission announced on the 2nd that it held an exemption review committee on the 30th of last month to deliberate and decide on such exemption measures.

Until now, the financial industry has continued to raise opinions that financial companies need to take measures against minor computer failures in this process in order to take active measures in response to the threat of frontier AI such as Mitos. There is a risk of computer system malfunction in a series of processes, such as checking attack surfaces using AI for security purposes, checking vulnerabilities, and executing security patches, because financial executives and employees are conscious of these risks and have limitations in taking active action against the immediate frontier AI threat.

In response, the Financial Services Commission decided to exempt executives and employees from sanctions such as imposing fines even if minor computer failures occur in testing AI for security purposes or executing security patches related to security vulnerabilities spread by authorities. Here, minor computer failures mean cases that fall within certain standards, such as △ no intention and △ monetary damage of less than 100 million won △ system failures within up to 4 hours △ less than 10,000 leaks of customer information (excluding personal credit information). In addition, immunity is made only on the premise that prompt restoration and consumer protection measures have been taken.

On this day, the Financial Services Commission also distributed guidelines that contain guidelines for financial companies’ actions against AI security threats. The guidelines contain six measures to respond to AI security threats: △ Strengthen management responsibility for AI security threats △ Weakness and patch management △ Asset and supply network management △ AI-based defense automation △ Strengthen joint response and resilience △ Anti-infringement system.

An official from the Financial Services Commission said, “As the domestic and foreign situation related to the frontier AI security threat is rapidly changing, we focused on inducing financial companies to take aggressive measures to strengthen security,” adding, “We will continue to push for various policy tasks to support the financial sector’s AI transformation, such as easing regulations on network separation.”