Tuesday, December 3, 2024

Bank of England Cautions Against Complacency in AI Development

The Bank of England is starting a consortium aimed at bringing together private finance organizations and AI experts to explore how technology can benefit the finance sector while managing risks.

At a financial conference in Hong Kong, Sarah Breeden, the Bank’s deputy governor for financial stability, emphasized the need for regulation to keep pace with AI adoption. She mentioned that understanding AI’s benefits and the various strategies firms use to handle risks is crucial for maintaining financial stability. The regulator plans to share best practices and determine when new guidelines might be necessary. Breeden warned against complacency, noting past experiences where addressing risks retrospectively became challenging as technology scaled up.

The Bank of England, along with the Financial Conduct Authority (FCA), has been examining how UK financial services firms are integrating AI and machine learning. In their latest survey involving 120 firms, Breeden found that 75% are using some form of AI. This marked a 53% increase since the previous year’s survey, which included all major UK and international banks, insurers, and asset managers.

She highlighted that 17% of these firms are utilizing foundation models, such as large language models like GPT-4. Many initial AI applications, according to the survey, are relatively low-risk regarding financial stability. For instance, 41% of firms are using AI to streamline internal processes, while 26% enhance customer support.

Looking ahead, Breeden noted that firms are increasingly turning to AI to address external risks, with 37% employing it to combat cyber threats, 33% for fraud prevention, and 20% for anti-money laundering efforts. There’s a shift in use cases, as 16% currently use AI for credit risk assessments, and 19% plan to in the next three years. Additionally, 11% are engaged in algorithmic trading, with another 9% intending to adopt this approach soon.

Breeden acknowledged AI’s potential to boost productivity and growth across the financial sector and the broader economy. However, she stressed that regulators need effective policy frameworks to manage the financial stability risks associated with AI to ensure economic stability.

She also pointed out two critical areas to monitor regarding generative AI in finance. First, at the individual firm level, regulators must ensure that technology-neutral frameworks can adequately address the risks that come with increasingly powerful AI models. Breeden emphasized the importance of financial firm managers understanding and managing their evolving AI systems. Second, she urged ongoing evaluation of regulatory boundaries, particularly if the financial system becomes more reliant on shared AI technologies and infrastructure.