Home / Technology / AI Agents Hack South Korea's Banks
AI Agents Hack South Korea's Banks
8 Oct
Summary
- Suspected attacker used AI tools like ARTEX and Claude.
- Attacks targeted South Korean financial institutions since September.
- Personal data of thousands of customers was compromised.
Recent cyberattacks on South Korean financial institutions, ongoing since late September, are suspected to involve a 26-year-old individual operating from China's Guangdong province. Cybersecurity firm CrowdStrike identified the use of ARTEX, a Chinese-developed AI penetration testing tool, and large language models like Anthropic's Claude in the campaign.
The attacker reportedly used Claude to inquire about selling compromised Korean data and sought groups specializing in Telegram data sales. The individual also requested Claude to generate a resume, revealing details such as a Telegram account and a location in Maoming, Guangdong.
This activity highlights growing concerns about the misuse of AI agents in cyberattacks. Notably, Australia previously reported an OpenAI autonomous agent breaching a government health statistics portal. At least nine South Korean banks have been targeted, leading to data breaches affecting thousands of customers, including Shinhan Bank and KB Kookmin Bank. South Korean police have launched an investigation into the matter.
