Bank of England warns banks about AI cyber risks

Speaking at the Institute of International Finance’s annual meeting in Bangkok, Bank of England Deputy Governor Sarah Breeden said advances in AI mean the technology now poses a systemic risk to the world financial system.[1] She said banks should not assume cyber attackers will fail to penetrate their systems.[1] Breeden identified hackers or hostile nations weaponizing advanced AI models to exploit weaknesses in banks’ systems as a chief concern.[1] PYMNTS also reported regulators’ concern that rogue AI agents could act unexpectedly and bypass banks’ safeguards.[1]

The Bank of England, Financial Conduct Authority and HM Treasury said current frontier AI models can exceed the cyber capabilities of a skilled practitioner at greater speed, scale and lower cost.[3] They said malicious use could threaten firms, customers, market integrity and financial stability.[3] The Bank said frontier AI may make cyberattacks faster and easier, outages more disruptive, and criminal scams more convincing.[2] The authorities’ warning therefore covers both attacks on firms’ systems and downstream harm to customers and the broader financial system.[2][3]

The authorities said firms should be able to protect against, detect, contain and respond to faster and more disruptive frontier-AI-driven attacks.[3] The Bank said it requires banks to demonstrate resilience through stress tests and penetration testing.[2] Separately, Breeden said the Bank is surveying banks about preparedness for quantum-computing risks to encryption, and will use the findings to develop a system-wide response for the financial sector.[1] The Bank has also called for stronger international coordination on testing frontier AI models before wider deployment.[2] Its note said it was not intended to introduce new expectations, but reinforced existing messages.[3]