South Korea Investigates Leak of 70,000 Bank Records Suspecting AI Use

South Korean police opened a wide-ranging investigation following the leak of around 68,000 to 70,000 customer records from seven financial institutions, according to a report by Vietnam.vn. The leak included data of about 25,000 customers at Shinhan, 40,000 at Yegaram Savings, and 2,200 corporate customers at Welcomb Savings Bank, in addition to 119 at KB Kookmin and 89 at Hana.

South Korean President Lee Jae-myung stated during a cabinet meeting that some of the attacks show signs of artificial intelligence usage, raising widespread concern, and called for intensifying resources to minimize damage. The Financial Services Commission had held an emergency meeting on October 4 after confirming that major banks had been targeted in the attack.

Server logs indicate that the attacker used an AI-powered penetration testing tool to automate vulnerability scanning and select infiltration methods, while Woori and NH NongHyup were able to detect abnormal access and block it in time. The leaked data includes names, phone numbers, annual incomes, and credit limits.

A professor at Chungnam National University noted that artificial intelligence "reveals vulnerabilities that might escape the human eye," urging proactive vulnerability detection and mitigation using the same technical models. The incident comes at a time when reports are growing regarding the use of AI agents in breaches targeting government entities in Australia and Canada, prompting financial institutions to review their defenses.

What do these terms mean?

Penetration testing tool: Software that simulates an attack on a system to discover its weaknesses before a real attacker exploits them. AI Agent: An artificial intelligence system that performs multiple steps autonomously to achieve a specific goal, rather than merely answering a question. Credit limit: The maximum amount a financial institution allows a customer to use via a loan or card.