• FSB chair Andrew Bailey warns frontier AI could amplify systemic cybersecurity risk
• Financial firms should prepare for simultaneous disruption across institutions and shared technology providers
• Bailey wants stronger controls governing frontier AI development and deployment
• Financial institutions are being urged to consider disconnected bare-metal backup systems
• AI risk is emerging alongside stretched equity valuations, rising leverage and private credit concerns
• The FSB is examining safe AI deployment and stronger cyber resilience

Artificial intelligence could magnify cybersecurity threats into a source of systemic financial instability, with financial institutions facing the prospect of simultaneous disruption across firms and critical technology providers, the head of the Financial Stability Board has warned.

Andrew Bailey, who also serves as Bank of England governor, urged financial services companies and technology providers to prepare for “more severe scenarios involving simultaneous disruption across multiple firms or shared technology dependencies.”

In a letter to G20 finance ministers and central bank governors, Bailey warned that recent developments involving advanced AI had exposed weaknesses in how some jurisdictions oversee the development, release and deployment of frontier models.

Concerns have intensified following testing involving flagship models developed by Anthropic and OpenAI, during which AI systems reportedly circumvented controls, accessed external organizations and created fake identities while pursuing assigned objectives.

“Recent developments have also highlighted to me that many jurisdictions do not have the protocols in place to manage the development, release and deployment of advanced frontier AI models, heightening risks for the financial sector and beyond,” Bailey wrote.

He called for more countries to take appropriate measures governing the release of frontier models, potentially putting the FSB’s position at odds with the lighter-touch approach adopted by the Trump administration.

President Donald Trump recently signed an executive order establishing a voluntary framework through which government agencies can inspect frontier models before deployment, but stopping short of giving authorities powers to prevent their release.

The White House had previously imposed export restrictions on Anthropic’s frontier models amid concerns they could be used to exploit cybersecurity weaknesses in critical infrastructure.

Those restrictions were removed in July after the company agreed to additional safeguards.

Bailey also urged financial institutions to consider “bare metal” backup infrastructure that remains completely disconnected from primary networks. Such systems could provide a route to restoring operations following a sufficiently severe cyberattack that compromises conventional connected backups.

The FSB is examining both how financial institutions can deploy AI safely and how firms can improve resilience against AI-enabled cybersecurity threats.

Those risks are emerging against an already challenging financial backdrop. Bailey identified energy-related inflationary pressures, higher interest rates, growing investor leverage and elevated equity valuations among vulnerabilities confronting the financial system.

“Markets remain vulnerable to a potentially disorderly correction that could spread across borders,” he warned.

Bailey also highlighted fragilities in sovereign debt markets, potential problems within private credit and equity valuations partly supported by enthusiasm about AI’s economic potential.

Growing leverage among equity investors represents another vulnerability. Bailey said increasing debt use was characteristic of a financial cycle reaching greater maturity, potentially magnifying market movements if sentiment deteriorates.

The warning places AI at both sides of the emerging risk equation. Investor expectations surrounding the technology are contributing to elevated market valuations, while increasingly capable models could simultaneously strengthen the tools available to cyber attackers.

For financial institutions, that combination raises the possibility that AI-related disruption could extend beyond individual operational incidents.

Shared technology dependencies and interconnected markets mean a sufficiently severe event could affect multiple institutions simultaneously, increasing the potential for operational disruption to become a broader financial stability problem.