Lawmaker urges government to accelerate AI hacking defenses as financial sector threats grow

Rep. Lee Hoon-ki of the Democratic Party of Korea, who represents Namdong-gu B in Incheon and sits on the National Assembly's Science, Technology, Information, Broadcasting and Communications Committee, called out the government's lack of preparedness for AI-driven hacking threats spreading through the financial sector at a national audit of ICT public institutions Thursday. He urged the Ministry of Science and ICT to act quickly.

A series of hacking incidents and personal data breaches has recently hit major banks, savings banks and capital companies, according to the financial sector. Traces of ARTEX — an AI-based autonomous penetration-testing tool — were found in some of the attacks, raising concerns about the potential misuse of AI.

At the audit, Lee questioned Choi Woo-hyeok, director general for information security and network policy at the Ministry of Science and ICT, about the differences between AI-driven hacking and conventional hacking, and about the government's readiness to respond.

"AI hacking attacks are characterized by being carried out automatically and simultaneously in a short period of time," Choi said. "However, even experts find it difficult to determine whether AI was actually used. In the case of the recent financial sector attacks, AI involvement is being inferred based on traces of ARTEX." He added that the ministry is pursuing a combination of basic security measures and AI-assisted vulnerability detection as countermeasures.

Lee also examined the Korea Internet & Security Agency's schedule and effectiveness in building an AI-based incident response system. The project received 5 billion won ($3.74 million) from last year's second supplementary budget and 15 billion won from this year's budget. Work on procuring equipment such as GPUs and developing the system is currently underway, with an additional 10 billion won earmarked for next year to complete the first phase of construction and conduct testing and verification.

"AI hacking appears to be just getting started and is expected to expand, so it is a problem that the AI hacking response system will still be at a testing stage even in the first half of next year," Lee said. "We need to hurry to prepare so we can respond to the AI hacking threats that have emerged in the financial sector."