Regulators said advanced AI systems now exceed skilled human practitioners in speed, scale and cyber capability
The Financial Conduct Authority (FCA), Bank of England and HM Treasury have issued a joint warning to financial firms over the cyber risks posed by rapidly advancing frontier artificial intelligence models.
In a statement published on 15 May 2026, the regulators said frontier AI models represent a “step-change in capability” with major implications for cyber security and operational resilience across the financial sector. They warned that the cyber capabilities of current frontier AI systems are already exceeding those of skilled human practitioners in terms of speed, scale and cost.
The authorities said these capabilities could significantly strengthen cyber defence measures if used responsibly, but warned they could also amplify cyber threats if deployed maliciously. According to the statement, the risks could affect firms’ operational stability, customer protection, market integrity and wider financial stability.
The joint intervention urged regulated firms to take proactive steps to assess and mitigate the risks associated with frontier AI technologies. Regulators said organisations should continue strengthening governance, cyber resilience and operational controls as AI systems become more sophisticated.
The statement follows increasing international concern over advanced AI models capable of identifying software vulnerabilities and automating elements of cyber intrusion. Recent debate within the financial sector has focused particularly on the implications of Anthropic’s Claude Mythos Preview model, which cybersecurity experts have warned could accelerate complex cyberattacks against banking systems and legacy infrastructure.
UK authorities said firms should also consider how AI tools can improve defensive capabilities, including threat detection and automated response measures. The regulators stressed that operational resilience frameworks must continue evolving alongside developments in frontier AI technology.
The statement forms part of wider regulatory activity on artificial intelligence within financial services. Earlier this year, the Bank of England confirmed plans to examine the impact of AI agents in financial markets, while the FCA outlined proposals to expand AI testing and oversight initiatives.
The FCA, Bank of England and Treasury said they would continue engaging with firms and international partners as frontier AI capabilities develop and associated cyber risks evolve.