The head of the Bank of England’s regulatory division has warned that the newest generation of artificial intelligence models could create “quite significant disruption” across the financial sector, especially through cybersecurity vulnerabilities and operational failures.
According to reports from Reuters, regulators are increasingly concerned about highly advanced AI systems such as Anthropic’s “Mythos” and newer ChatGPT models, which are capable of identifying software weaknesses at unprecedented speed. Officials fear that banks and financial institutions may struggle to patch vulnerabilities quickly enough, potentially leading to outages or systemic disruptions.
The concerns come as UK authorities — including the Financial Conduct Authority and Britain’s National Cyber Security Centre — have reportedly been holding urgent discussions with banks, insurers, and exchanges to assess the risks posed by frontier AI systems.
Regulators say the danger is not only from malicious hackers using AI, but also from the growing dependence of financial institutions on the same AI tools and cloud systems. If many firms rely on similar models, a single vulnerability could trigger widespread failures across the banking system.
The warning follows broader international concern. The International Monetary Fund recently cautioned that advanced AI models could create “systemic” financial risks by accelerating cyberattacks and exposing hidden weaknesses in critical infrastructure. Meanwhile, the European Central Bank has also begun studying defenses against AI-powered attacks.
Anthropic’s experimental “Mythos” model reportedly discovered thousands of serious vulnerabilities in operating systems, browsers, and commonly used software during testing. While this capability could improve cyber defense, regulators worry the same technology could be exploited by criminals or hostile states.
Bank of England Governor Andrew Bailey previously said the technology may have “cracked the whole cyber risk world open,” highlighting the scale of concern among policymakers and central banks.
