South Korean authorities are investigating several cyberattacks in which attackers reportedly used artificial intelligence (AI) to target customer data at major banks. The incidents involve Hana Bank, KB Kookmin Bank and Shinhan Bank.
Scale of the data exposures
According to reporting by The New York Times, at least 25,000 customers' loan records from Shinhan Bank were leaked during the past week. KB Kookmin Bank was affected in the case of 99 customers and 20 current or former employees, while Hana Bank reported the theft of data belonging to 89 customers.
Official and political responses
The cases are under investigation by South Korean authorities. President Li Dzsemjong said available signs indicate the attackers used AI models, a development that has "provoked significant public outrage and concern," and he instructed the government to clarify the situation and take all possible steps to minimize harm.
Why AI is a concern for cybersecurity
Reports note that AI models become more capable with each release, and malicious actors are leveraging those capabilities to probe weaknesses in security systems. Anthropic reported in November 2025 the first cyberattack carried out with AI assistance by a Chinese hacking group. A few months later, Taiwan was reportedly the first target of a fully automated attack by the same actors.
Experts warn that attackers can exploit AI models' so‑called "hallucinations" — inaccurate or unintended behaviors — to coerce agents into producing or running malicious code. At the same time, leading AI companies say they build defensive mechanisms into their systems to block illegal or unethical uses of their models.
Human factor remains relevant; perpetrators unidentified
Specialists emphasize that the human element cannot yet be ruled out in these attacks, but AI lowers the technical barrier to conducting them: attackers need not be experts in the subject to mount such incidents. Authorities have not yet publicly identified who is responsible for the recent breaches.
What comes next
Investigations are ongoing. Authorities and the affected banks must now inform impacted customers, assess the damages and prevent further data loss. The incidents also underscore a broader, global challenge: advancing AI capabilities create new risks for the financial sector that will require regulatory, technical and operational responses.
