The AI-assisted raid on South Korea's banks never went near the vault. It slipped in through loan portals, sales platforms and a staff mobile app. Now Seoul is ordering every bank to count its exposed doors, and regulators in Washington, London, Frankfurt and Sydney are asking the same question.
A bipartisan AI committee wants independent evaluators inside OpenAI and Anthropic and early access for security agencies. Read slowly, its letter to Albanese is a ransom note written in reverse. Australia is not being held hostage. It is offering itself, with its creative archive as collateral.
Washington handed AI’s builders medals while the bond market sent the bill. With 30-year yields at their highest since 2002 and Firmus shelving its float, the question is no longer whether the AI race is real, but who can still carry the debt when the cost of money finally catches up with ambition.
The AI-assisted raid on South Korea's banks never went near the vault. It slipped in through loan portals, sales platforms and a staff mobile app. Now Seoul is ordering every bank to count its exposed doors, and regulators in Washington, London, Frankfurt and Sydney are asking the same question.
South Korea’s bank raids came through systems nobody was watching. From Seoul to Sydney, supervisors are now counting every door.
The intruder who worked through South Korea’s financial sector over the past fortnight never went near the vault.
Korean authorities traced the same attacker address across seven financial firms, and each time the target was the soft periphery rather than the hardened core: a service used by loan agents, sales-support platforms, a mobile tool for staff. Shinhan Bank alone lost data on about 25,000 customers.
The weapon was as unglamorous as the targets. CrowdStrike has linked the campaign, with moderate confidence, to a 26-year-old in Guangdong who denies involvement, running ARTEX, an open-source Chinese penetration-testing agent, alongside Anthropic’s Claude Code. ARTEX is a harness rather than a model, built to borrow whichever language model will cooperate.
That is the quiet lesson of this year’s jailbreak debate. Safeguards live inside models while attacks increasingly live in the software around them, and even the models are porous: Britain’s AI Security Institute found universal jailbreaks in OpenAI’s GPT-5.6 Sol, often within hours.
Seoul’s response says more than any communique. The Financial Services Commission has ordered banks and card issuers to inventory every IT asset reachable from outside their networks, close any unauthenticated path to internal data and report back, while inspectors from three agencies work on site at Shinhan, KB Kookmin, Hana and BNK Busan.
“We need to check the entire security framework to ensure there are no gaps.” - Lee Eok-won, Chairman, Financial Services Commission, South Korea
Western supervisors are asking the same questions. Reuters reported this week that the Federal Reserve, the OCC and the FDIC are quizzing banks on whether they run Chinese-developed technology, use AI agents in cyber defence and can shut autonomous systems down.
The Bank of England and the European Central Bank are pressing lenders on recovery, with British supervisors encouraging the largest to examine bare metal systems that can run independently of a compromised estate. The ECB’s deadline for AI-cyber action plans falls at the end of October.
“The adoption of AI is often outpacing the ability to put the controls in place.” - Michael Woodbridge, Cybersecurity Partner, Grant Thornton
Australia’s regulators have been making the same argument since April, when APRA warned that frontier models would raise the probability, speed and scale of attacks on the financial system and later urged the major banks to share tools and intelligence with smaller lenders. The Australian Signals Directorate has gone further, telling enterprises that prompt injection has no fully reliable fix inside the model and that controls belong in the harness around it.
From Seoul to Sydney the pattern holds. Banks spent a decade fortifying the vault while leaving the side doors to vendors, contractors and convenience, and AI has made it cheap to try every handle at once.
The institutions that come through this season intact will be those that know where all their doors are, can close any of them in an afternoon, and can rebuild the house quickly if someone walks in regardless.
Get the stories that matter to you.
Subscribe to Cyber News Centre and update your preferences to follow our Daily 4min Cyber Update, Innovative AI Startups, The AI Diplomat series, or the main Cyber News Centre newsletter — featuring in-depth analysis on major cyber incidents, tech breakthroughs, global policy, and AI developments.
OpenAI flew its chief strategy officer to Sydney to apologise for a rogue AI agent's Medicare breach. The hack was clumsy. The silence that followed was worse, and Parliament now has to decide whether sorry is the only accountability on offer.
OpenAI’s Medicare portal breach has become a test of AI accountability. An agent crossed an access boundary while pursuing public health data, yet Australia was notified months later. No records appear to have been accessed, but the incident raises questions on safeguards and trust.
From the UN in New York to a proposed AI campus near Dalby, Australia is seeking a voice in rules and a stake in the infrastructure. Albanese’s diplomatic push raises a practical question: can global ambition deliver local benefits while protecting energy, water and Australia’s digital sovereignty?
Dario Amodei wants a speed limit for frontier AI. Elon Musk and Sam Altman back the direction, but Washington says the labs should brake first. As China shapes the race, Cyber News Centre examines whether independent oversight can turn safety promises into controls before a new agent swarm escapes.
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead. Sign up for free!
