Australian banks warned frontier AI could create larger, faster cyber attacks

Australia’s prudential regulator APRA warned on April 30, 2026 that banks are not keeping pace with frontier AI developments, saying systems like Anthropic’s Claude Mythos could accelerate and scale cyber attacks by helping bad actors discover vulnerabilities. APRA flagged that many banks rely on vendor model presentations, have limited board technical literacy and need a “step change” in cyber practices. The story notes banks already spend “billions each year” on security and regulators are engaging with firms including Anthropic. Ratings agency S&P Global warned AI may affect credit standings of Asia-Pacific financial institutions over the next one to five years, though large tech budgets could mitigate impacts. Market implication: elevated operational and credit risk for banks could pressure financial stocks and indices without stronger controls, while AI also offers cost and efficiency benefits that may offset some negatives.