European regulators warn banks on AI cyber threats

European and UK regulators have issued urgent warnings about the financial stability risks posed by advanced artificial intelligence, particularly frontier models like Anthropic's Mythos. The European Central Bank (ECB) has given eurozone banks until October 31 to submit plans to counter AI-enabled cyber attacks, citing models that can identify software vulnerabilities and write exploits at "unprecedented speed" Source: thenextweb. The European Systemic Risk Board (ESRB) has raised its systemic cyber risk assessment to "severe" Source: euronewsen.

Bank of England (BoE) actions

The BoE separately highlighted AI as a growing threat to financial stability, noting that rapid AI advances increase cyber risks and could amplify market volatility. The BoE is also considering loosening capital rules for UK lenders, a move some FPC members worry could increase debt-fuelled investments in AI-related stocks Source: theguardian.

Key concerns across jurisdictions

• Cyber resilience: Both regulators stress that AI enables faster, larger-scale attacks; the BoE warns that defenders may struggle to keep pace Source: cna.
• Market risks: The BoE and Bank for International Settlements warn that a reassessment of AI earnings prospects could trigger sharp asset price falls Source: cointelegraph.
• Regulatory challenges: UK FCA CEO Nikhil Rathi said traditional rulemaking "simply doesn't work" with AI moving in weeks Source: cointelegraph.

Divergent policy approaches

While the ECB imposes a hard deadline for bank plans, the BoE proceeds with capital easing despite internal dissent. Both acknowledge that frontier AI developers sit mostly outside Europe, creating strategic dependency Source: thenextweb, euronewsen.