South Korea believes AI agents were used to hack several banks

[Paige_26]

South Korean officials believe AI agents were used in breaches at seven or more financial institutions, exposing personal data of at least 68,000 people, The Record reports. Officials believe a Chinese cybersecurity tool called Artex AI was used. Financial authorities have identified 33 IP addresses involved in the attacks, traced to at least 12 countries including Japan, the US, Thailand, Vietnam and Hong Kong. The Record calls it the first known case of AI agents hacking the financial sector

Hana Bank, KB Kookmin Bank and Shinhan Bank were among those hit. Shinhan reportedly said data on about 25,000 customers was exposed, including borrowing history, incomes, phone numbers and names. The incidents came to light on 30 September. The National Office of Investigation has set up a 28 person team to look into the attacks

President Lee Jae Myung said "Signs have emerged" that AI agents were used in at least some of the attacks. He warned that it is now possible to hack "with ease even without specialized skills." That is the worry security experts have raised for some time. AI agents could lower the skill needed to carry out serious attacks, putting powerful tools in the hands of far more people

The article links the case to Australian officials' claim last month that AI agents hacked a Medicare database, and to the Hugging Face incident involving OpenAI's models that we have discussed several times. Taken together, the stories suggest AI driven attacks are moving from theory to reality. Banks and other organisations will need defences that can keep up

Are AI powered cyberattacks the biggest new threat to banks? And what should ordinary customers do to protect themselves?

South Korean officials believe AI agents were used to hack several bankstherecord.media

[SingularityNodeOwl]

Borrowing history and income data being exposed is nasty. That is exactly the kind of information scammers use to make convincing calls pretending to be the bank

Leaked details make scams sound far more believable. Banks should warn affected customers to be extra careful with calls and texts for a while

[Inland Aidan]

With the Hugging Face incident, the Australian claims and now this, it has been a worrying few months. Regulators need to take AI security seriously. Every month seems to bring a new AI security story. It is no longer a theoretical risk

[Freddie_85]

Defenders will use AI too. Banks already use machine learning for fraud detection. It will become an arms race between attacking and defending AI systems. The question is whether defence can keep pace with attackers who only need to succeed once. Neither side can afford to stand still

[GhostRider89]

Customers cannot do much about a bank being breached. But turning on alerts for every transaction and watching for scam calls that use leaked details helps limit the damage. Never give out codes or passwords over the phone, whoever the caller claims to be. Banking apps with instant notifications make spotting fraud much quicker

[Sentinel66]

This is exactly what security researchers warned about. AI agents that can scan for weaknesses and exploit them at speed change the economics of hacking. Attacks that needed a skilled team can now be run by one person. Agents that can try thousands of approaches without tiring are a nightmare for defenders

[Carol_83]

Seven institutions at once suggests automation. Human hackers usually focus on one target at a time. Automated attacks can probe hundreds of targets at once

Banks need to assume they are always being tested. Speed is the thing that worries me most