From Startup Fortune · · 1m 20s
South Korean Regulators Confirm AI Agent Breached 16 Banks
South Korean regulators confirm an AI agent hacked 16 banks, exposing 100,000 records.
In brief
South Korean regulators confirm an AI agent hacked 16 banks, exposing 100,000 records. South Korean regulators confirmed an AI agent breached 16 banks, exposing 100,000 records including names, incomes, and national IDs. Originally reported by Startup Fortune.
Regulators Confirm 16 Banks Breached
South Korea's Financial Services Commission confirmed 16 financial firms were breached since October began.
AI Agent Used in Attacks
A Chinese security engineer posted an open-source AI pentesting agent to GitHub on July 26.
100,000 Records Exposed
The breach exposed names, incomes, loan limits, and national ID numbers across 100,000 records.
28 Cyberterrorism Investigators
Police assigned 28 cyberterrorism investigators across four teams to determine the attack's depth.
ARTEX AI Tool Details
ARTEX, an open-source AI penetration-testing agent, identifies targets, finds vulnerabilities, and plans attacks without human steering.
Attack Targeted Peripheral Systems
Attackers targeted loan-agent portals and mobile employee tools, not the banks' core ledgers.
26-Year-Old Chinese Hacker
CrowdStrike traced the attacks to a suspected 26-year-old in Guangdong province, China.
AI Models Used in Attacks
The attacker used DeepSeek v4.1-flash, Zhipu AI's GLM-5.3, and xAI's Grok 4.6 in the breach.
ARTEX Pulled from GitHub
The ARTEX developer pulled the tool from GitHub on October 8 after breaches were linked to it.
The bottom line
South Korean regulators confirmed an AI agent breached 16 banks, exposing 100,000 records including names, incomes, and national IDs.
