Regulators Flag Systemic Concerns about the Risks of AI in Finance

As artificial intelligence (AI) becomes more embedded in financial systems, regulators are increasing their scrutiny of the inherent risks that come with growing reliance on AI-powered automation. Their concerns were highlighted in a recent report published by the Bank of England, which outlined several key areas that require close monitoring. A primary concern focuses on the use of AI in core financial decision-making, where complexity and a potential lack of transparency—or ‘explainability’—in many AI solutions could lead to serious problems, such as biased or discriminatory decision-making or taking on poorly understood risks.

Regulators also fear that the widespread use of AI in financial markets could lead to unexpected and undesirable behaviors, particularly with regards to trading and investment markets. Already, a great deal of trading is carried out algorithmically, but AI has now reached the point where it can be difficult to tell the different between an AI-generated decision and a human one. This could potentially amplify market volatility, especially if AI systems are allowed to make decisions unchecked—a serious concern given the imminent rise of fully automated agentic AI systems.

Operational risk is another major focus area, given the financial and fintech sectors’ growing reliance on a relatively small number of third-party providers—like Google, Microsoft, and OpenAI—for critical services. The high concentration of risk in such a small number of companies—almost all of which are headquartered in the US—could lead to very undesirable global systemic consequences. The Financial Stability Board (FSB) echoes these concerns, particularly regarding provider reliance and AI-enhanced fraud and other cyber threats.

Proactively managing the risks of AI

As regulators struggle to keep up with the inexorable march of AI-centered innovation, it might seem tempting for fast-growing fintechs to innovate rapidly, bolting on regulatory compliance and security as something of an afterthought. However, both the short- and long-term risks of this approach are immense. For instance, a lack of sufficient model validation, testing, and ongoing monitoring can result in serious consequences to brand trust and compliance with current and future regulatory demands.

To counter these very real risks, fintechs and financial services firms must continuously assess third-party AI risk, deploy and enhance AI-specific cybersecurity measures, and monitor regulatory guidance to ensure their innovation efforts are sustainable and don’t add business risk. Regulators might take some time to catch up with the new and emerging threats that come with AI, but fintechs can’t afford to rest on their laurels when their entire industry is built on trust and transparency.