Suspected AI-agent cyberattacks affect 16 financial firms

The reports described approximately 100,000 pieces of information leaked from 16 financial institutions, including Shinhan, KB, Hana and Busan Bank.[1] The exposed information included names, contact details and resident registration numbers, while regulators found no evidence that account or card numbers were compromised.[2] The affected firms spanned banks, savings banks, credit finance companies, electronic payment providers and cryptocurrency businesses.[2] Traces of ARTEX, an AI autonomous penetration-testing tool, were found at the attacking IP address.[1] That trace is evidence of possible tool use, not confirmation that AI carried out the attacks.[1]

The reporting said AI tools designed to check security vulnerabilities may have been abused in actual attacks.[1] AI-utilizing attacks can be automated, carried out in a very short time and launched across multiple fronts simultaneously.[1] At the same time, even experts may find it difficult to confirm whether an attack was caused by AI.[1] Investigators were examining whether the attacks were carried out by the same hacking group.[2] For security teams, the combination makes it important to investigate the activity without treating the presence of an AI tool as conclusive proof of AI-led responsibility.[1]

Regulators found security deficiencies at 130 of 555 financial companies inspected through October 8, equal to 23.4 percent.[2] Reported problems included missing updates, weak oversight of unsupported systems, no 24-hour monitoring, and gaps in external-system authentication and authorization.[2] The Financial Services Commission planned sector-wide security reviews and broader reforms, including greater accountability for CEOs and boards.[2] The report characterized continued failures to meet basic security requirements as deeply troubling despite repeated calls for inspections and stronger controls.[2]

A pilot project is running to let white-hat hackers find and quickly fix vulnerabilities through attacks.[1] Under the current Information and Communications Network Act, white-hat hackers attempting attacks are committing illegal activity.[1] The government plans to pursue institutional changes related to the program after the pilot.[1] A separate security-focused AI foundation-model project, called “Teuk-pa-mo,” began in September and is targeting use next year.[1] The reporting also described discussions about integrating AI into existing security systems to find and fix problems before attackers can exploit vulnerabilities.[1]