GenAI and Financial Risk Management

The 2008 financial crisis was, above all, a massive failure of risk management. Since then, financial institutions have been forced to raise their game but (as the case of Credit Suisse and Silicon Valley Bank has shown) there are still lessons to be learned.

In the meantime, however, the tools and technologies available to financial institutions have been changing rapidly. Like many other industries, the financial sector is now looking to generative artificial intelligence (GenAI) as a potential solution to its problems, including risk management. In this article, we explore the potential benefits and challenges of using GenAI as a risk management tool.

First steps for AI models in risk management

Financial institutions are usually late adopters. Highly regulated industries tend to be wary of new tools, even as the rest of the world rushes to embrace them. The Global Association of Risk Professionals (GARP) echoes this sentiment but also says that’s beginning to change: “However, the advent of GenAI presents an unprecedented opportunity to supercharge risk management operations, making them not just more effective but also significantly more efficient. For visionary firms, instead of just playing a supporting role, this disruptive technology can take center stage.”

The financial services sector is not new to the application of AI technologies. In fact, for some time AI has been used to tackle repetitive and data-heavy tasks. For instance, as far back as 2018, Jeanne Boillet wrote for EY, “Machine learning can support more informed predictions about the likelihood of an individual or organization defaulting on a loan or a payment, and it can be used to build variable revenue forecasting models.”

Additionally, machine learning algorithms are used to detect credit card fraud and block suspicious activity. “Financial institutions also use automated systems to monitor their traders by linking trading information with other behavioral information such as email traffic, calendar items, office building check-in and check-out times, and even telephone calls,” according to EY.

GenAI as threat detector and compliance assistant

The advent of GenAI has greatly widened the field of potential application. GARP points out that GenAI 's massive information processing potential is a game-changer: “This capability offers immense potential for risk managers, who must constantly monitor various internal and external data sources to detect and rank threats to a company's operations or investment portfolios.”

McKinsey suggests GenAI can help keep employees on top of regulatory compliance. Companies can train bots to answer questions about regulations, company policies, and guidelines. “As a code accelerator, it can check code for compliance misalignment and gaps,” McKinsey reports. “It can automate checking of regulatory compliance and provide alerts for potential breaches.”

The opportunities for GenAI to create efficiencies are only growing, but every new tool comes with a warning.

Potential pitfalls and challenges of GenAI adoption

Despite its promise to revolutionize risk management, GenAI must be used with caution. First and foremost, it is imperative to remember that algorithms can reinforce existing biases. The Associated Press (AP) reported on an Investigation from The Markup that found that due to algorithmic biases, lenders were 80% more likely to reject Black applicants, 70% more likely to deny Native American applicants, 50% more likely to turn down Asian/Pacific Islander applicants, and 40% more likely to reject Latino applicants.

AI is fallible in other ways. It is susceptible to poor-quality or incomplete data and can suffer from programming errors. In spite of these weaknesses its use is growing fast: Interviewed in the Financial Times, analyst Andrew Schwartz estimated that "more than half of large financial institutions are, at present, using AI to manage risk."

AI also opens users up to potential liabilities, especially as the legal landscape changes to keep up with this emerging technology. Speaking to The Financial Times, Zayed Al Jamil, a partner at law firm Clifford Chance, said of regulators, “They will not say that [AI] is banned [for risk management] or be extraordinarily prescriptive . . . I think that they will update existing regulations to take into account AI.”

The same article also points to the 'black box' problem. " ... the complexity of ChatGPT and similar AI technologies may make it hard for financial services firms to explain their systems’ decisions. Such systems, whose results are inexplicable, are known as “black boxes” in AI jargon."

A balanced approach to GenAI adoption

By leveraging GenAI technologies, financial institutions can unlock a range of potential benefits, including enhanced accuracy in risk assessment, improved decision-making processes, and greater efficiency in compliance management.

However, if there is one keyword that seems central to the conversation around GenAI for risk management, it is “caution.” Of course, in a highly regulated industry like financial services, caution is always warranted when dealing with new tools, but the promise of GenAI is too great to ignore. So, what’s a risk management professional to do?

Picking the lowest-stakes applications can help ease risk management pros into AI. For example, using a chatbot to quickly answer compliance questions is lower stakes than using AI to accept or reject mortgage loans. Putting AI to work in monitoring threats is a great way to enhance the capabilities of your cyber security team without running afoul of regulators.

Achieving a balance between the contribution of AI models and human judgement in risk management will be key: International Banker reminds us: “While it’s clear that AI will increasingly be used to automate a range of functions, ... there is inherent and irreplaceable value in human insight, judgement and decision-making, especially in areas as critical as risk management, where experience plays a massive role across the board.”