Andrew Bailey, Governor of the Bank of England and Chair of the Financial Stability Board (FSB), has warned that the rapid advance of frontier artificial intelligence (AI) models could become a serious threat to global financial stability, with cyberattacks emerging as the most immediate danger.
In a letter to G20 finance ministers and central bank governors, Bailey said increasingly capable AI systems were acquiring greater autonomy, sophisticated problem-solving abilities, and the capacity to identify and exploit software vulnerabilities. That could allow attacks to be carried out faster, at a greater scale, and at a lower cost, potentially disrupting several financial institutions at the same time.
The Financial Stability Board said frontier AI models had created across markets and financial centers a new risk landscape because financial institutions, markets, and critical infrastructure are highly interconnected and often rely on the same technology providers.
A successful AI-enabled attack on a shared supplier, software component, or critical service could therefore spread across multiple firms and jurisdictions. Bailey urged authorities to support the safe and responsible release and deployment of frontier models, while financial institutions should strengthen their ability to respond to and recover from major cyber incidents.
The Bank of England has already identified the issue as a financial stability concern. In its July Financial Stability Report, the central bank said rapid advances in frontier AI had significantly increased cyber and operational resilience risks. It noted that the latest models were increasingly able to conduct multi-stage attacks and exploit vulnerabilities with limited human input.
The Bank also said AI could accelerate the process of finding weaknesses, forcing financial firms to patch systems more frequently. That creates another operational risk: organisations could be pushed into making rapid changes to critical systems, increasing the possibility of disruption if those changes are poorly managed.
Bailey’s concerns extend beyond cyberattacks. He warned that financial markets could also be vulnerable to a disorderly correction in AI-related assets. The Bank has highlighted rising valuations and growing concentration in equity markets, where a relatively small group of technology companies has become increasingly important to major indices.
AI investment has created powerful links between chipmakers, cloud providers, data-centre operators, AI developers and technology companies. The Bank has warned that these relationships can create self-reinforcing capital loops. A sharp fall in expectations about AI profits could therefore spread losses across companies that are financially or commercially connected.
Leverage adds another layer of risk. Bailey has pointed to growing borrowing and leveraged investment in markets, which could amplify the effect of a sudden repricing of AI assets. If investors were forced to sell positions simultaneously, the resulting market moves could become more severe.
There is also a potential risk from the growing use of AI within financial institutions. Banks, insurers, and investment firms are increasingly using AI for research, coding, surveillance, customer service, and other functions. More autonomous systems could eventually play a larger role in financial decision-making and trading.
That prospect raises questions about whether firms and regulators will be able to understand and control AI behavior when market conditions change rapidly. If many institutions rely on similar models, they could respond to information in similar ways, potentially increasing correlation and reducing the diversity of decisions that normally helps markets absorb shocks.
The international nature of the technology makes coordination particularly difficult. AI models can be developed in one country, operated through infrastructure in another, and used to attack financial institutions across several jurisdictions. Differences in national regulation and cybersecurity capabilities could leave gaps that attackers exploit.
Bailey is therefore calling for a more coordinated global approach.
The FSB wants authorities to ensure that AI development and deployment to be accompanied by appropriate safeguards, while financial firms need stronger resilience and recovery capabilities. That includes the ability to restore critical systems after a major attack rather than assuming that existing digital infrastructure will remain available.
The warning does not mean Bailey sees AI only as a threat. The bank has acknowledged that the technology could generate significant productivity gains, transform financial services, and strengthen cyber defenses. The problem is that the same capabilities that help defenders identify vulnerabilities can also make attacks more sophisticated.
For financial markets, the challenge is to capture those benefits without allowing technological dependence to create a new source of systemic fragility. Bailey’s intervention puts that balance firmly on the G20 agenda at a time when investors are pouring unprecedented capital into artificial intelligence.
The message from the UK central bank is increasingly clear: AI risk is no longer simply a technology or cybersecurity issue. As advanced models become embedded in the financial system, failures, attacks, or market shocks involving them could quickly become financial stability events. The stakes are therefore enormous.