The Bank of England plans to revise its AI regulations in finance, citing significant risks associated with autonomous systems and cyber threats.
The Bank of England is set to review its regulatory framework concerning artificial intelligence in the financial industry, as stated by Deputy Governor Sarah Breeden during an event hosted by the European Central Bank in Sintra.
Breeden noted that existing technology-neutral regulations do not adequately cover systems that can function autonomously, highlighting significant risks tied to AI, such as cyber threats, payment failures, and potential market disruptions.
“The rapid advancements in AI, even just in the past six months, pose a dual challenge: ensuring responsible use and risk management while also transforming how central banks operate,” she remarked.
Potential new regulations may include more stringent recovery protocols for critical systems, enhanced backup capabilities, and rapid restoration processes for compromised infrastructure. Additionally, the Bank is considering implementing safeguards akin to emergency trading halts to mitigate market disruptions caused by faulty AI operations. Breeden also pointed out the risk of collective behavior among AI agents, which could heighten market volatility.
On May 15, the Bank, alongside the Financial Conduct Authority and HM Treasury, issued a joint statement addressing the cyber risks associated with advanced AI in finance. A subsequent warning from the Five Eyes alliance on June 22 noted that the gap between identifying vulnerabilities and their exploitation is narrowing.
Furthermore, the Financial Stability Board released a consultation document on June 10, detailing twelve practices for the responsible integration of AI in financial entities, aiming to enhance governance and risk management throughout the development and implementation processes.
According to research from the Cambridge Centre for Alternative Finance, a significant 81% of surveyed industry players are utilizing AI in some capacity, with 52% either in pilot stages or having advanced implementations. Most current applications focus on internal functions such as automation and data visualization, while front-office usage remains limited.
In an April 1 response to a parliamentary committee, the Bank of England indicated that generative and agent-based AI have not yet posed a systemic risk. Nonetheless, a reevaluation of the technology-neutral regulatory stance is forthcoming.
In May 2026, the IMF emphasized the need for AI-related cyber risks to be considered a matter of financial stability rather than just IT security.
Discussions regarding AI regulations are also taking place at other institutions, including the Bundesbank and the Bank for International Settlements.
