
Image to help understand the article
Customer Data Compromised
Shinhan Bank said personal information belonging to about 25,000 customers was exposed after an unauthorized outside user bypassed an identity-verification step in a service used by loan brokers.
The compromised records included customers’ names, phone numbers, annual income and calculated borrowing limits — information collected during the loan application process. The breach also exposed 66 South Korean resident registration numbers, which function similarly to Social Security numbers in the United States, and 97 connecting information, or CI, identifiers used to verify people across online services.
Shinhan Bank President Jung Sang-hyuk acknowledged the breach in a statement posted on the bank’s website and promised full compensation if customers suffer financial losses.
Signs of a Possible AI-Powered Tool
Cybersecurity specialists examining a web server believed to have been involved in the attack found a Chinese-language page title reading “ARTEX-自主渗透試控制台.” The wording roughly describes a console for autonomous AI penetration testing and could indicate that software called ARTEX AI, or a similar system, was operating on the server.
Moon Jong-hyun, head of the security center at South Korean cybersecurity company Genians, disclosed the finding on LinkedIn. He said threat analysts had reasonable grounds to suspect that an AI-based attack automation tool may have been used against the financial institution.
ARTEX AI is freely available on GitHub and aimed at Chinese-speaking users. It combines a large language model with multiple AI agents to automate tasks such as gathering information, identifying vulnerabilities, planning attack routes, operating security tools and verifying results. The software previously won an “Agent+” offense-and-defense competition organized by Baidu’s security response center.
Neither Shinhan Bank nor South Korean financial regulators have confirmed that ARTEX AI was used in the breach.
Open-Source Software Complicates Attribution
Security experts warned that tools designed for authorized security testing can also make real-world cyberattacks faster and more efficient when misused. Tasks that once required attackers to repeatedly enter usernames and passwords by hand can increasingly be automated across an entire attack sequence.
The Shinhan incident is believed to have involved credential stuffing, in which username-and-password combinations leaked elsewhere, including on the dark web, are tested across multiple websites. People who reuse the same login credentials for different services are particularly vulnerable to this type of attack.
Experts cautioned that the Chinese text does not establish who was behind the breach. Because ARTEX AI is open-source software that anyone can download without approval, the evidence cannot by itself identify the attackers or their location. The broader concern, they said, is that automated attacks using widely available software could become more common.
South Korea’s Financial Supervisory Service is conducting an on-site investigation after receiving Shinhan Bank’s incident report. Regulators are examining how the attacker gained access and how the customer data was removed.
Comments
Post a Comment