Central Banks in Australia and New Zealand Raise Concerns Over Mythos AI Risks

Global regulators are starting to react to Anthropic’s Mythos AI model, with central banks now stepping in to assess potential risks. Authorities in Australia and New Zealand have confirmed they are monitoring the situation closely, hinting at an unease around how advanced cybersecurity AI could reshape existing systems.

Central banks are reportedly concerned about Mythos rollout

According to a report by Reuters, the Reserve Bank of Australia is actively engaging with regulators, government bodies, and industry players as Mythos gains attention. Meanwhile, the Reserve Bank of New Zealand noted it is coordinating with both domestic and Australian counterparts over what it describes as a “developing risk.”

The concern stems from Mythos’ core capabilities. While designed for defensive cybersecurity tasks, Anthropic has already indicated that early testing uncovered thousands of major vulnerabilities across operating systems and web browsers. That alone appears to be raising concerns, especially as such tools could expose weaknesses faster than companies can respond.

Security experts have pointed out that Mythos may be capable of identifying and exploiting unknown vulnerabilities at a pace that outstrips traditional patching cycles. That could pressurize industries that depend on older or highly interconnected systems.

Banking is one area being closely watched. Many institutions still operate on legacy infrastructure, which could make them more vulnerable if tools like Mythos are used offensively. German central bank president Joachim Nagel described the model as a “double-edged sword,” highlighting its potential to strengthen security while also exposing systems to misuse.

Anthropic has released Mythos under a restricted program called Project Glasswing, granting access to select partners including major tech companies such as Amazon, Microsoft, Nvidia, and Apple. That said, a new Bloomberg report has claimed that Mythos AI has been access by unauthorized groups. Anthropic has confirmed that it is “investigating a report claiming unauthorized access to Claude Mythos Preview through one of our third-party vendor environments.”