Transforming Risk Management with Generative AI

The revolutionary potential of generative artificial intelligence applications, like ChatGPT or Google's Bard, is undeniable. Fueled by the promise of skyrocketing productivity and a turbocharged pace of research and product innovation, these so-called GenAI tools have sparked an intriguing dialogue among corporations and investors.

Can financial institutions harness this same power to redefine risk management?

Risk departments have traditionally focused on the prevention of potential pitfalls associated with adopting technologies such as AI, especially without robust safeguards. Their prime concern has been the protection of their organization's intellectual property and trade secrets.

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.

Let’s now examine three pivotal risk functions that stand to be revolutionized by GenAI. Instead of focusing on minimizing the risk of this technology, we can leverage it to reshape the entire risk management landscape.

Data Mining: Much More Than a Search Engine

At its essence, GenAI functions as a tool for processing information. Similar to search engines like Google and Bing, it excels at quickly finding pertinent information to resolve specific issues or queries. However, it surpasses these tools by providing curated, precise answers instead of a mere list of links for users to review.

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. The diverse nature of these threats often makes it challenging for a team of risk managers, let alone individual ones at smaller banks or credit unions, to stay current.

Consider a practical application where a GenAI tool constantly monitors data streams on recognized cyber threats, cross-referencing them with an institution's system profile to pinpoint specific vulnerabilities. After identifying a threat, the system could automatically alert risk managers and other relevant individuals in real time.

The true power of the AI system would be to then proactively source patches for these threats directly from approved software vendors for system engineers to implement. This would not only streamline the threat identification process but also enable the risk team to preemptively coordinate a solution, rather than delegating the problem to another team.

GenAI's capacity to automate the monitoring and processing of a company’s own internal data could make an even more significant contribution. Equipped with the latest statistical tools and algorithms, risk managers could leverage AI to compile and scrutinize transaction data across various departments for anomalies or outliers linked to specific suppliers or operations.

The speed and capacity to integrate multiple information sources can dramatically enhance the productivity of individuals who might otherwise rely on spreadsheets to assess outdated data. (My company, Moody’s, recently deployed a copilot tool to all of its employees that allows us to collate and analyze information from our massive trove of data across various risk categories.)

Democratization of Data and Analytics

The greatest opportunity for GenAI to revolutionize risk management may be in the democratization of access to risk data. This technology allows individuals without specialized technical skill sets to extract meaningful insights from company databases. Instead of needing proficiency in SQL or Python, GenAI enables users to query data using everyday language, simplifying tasks that previously required extensive programming knowledge and time.

Given the ease of use, GenAI can empower everyone within an organization to contribute to risk management. For example, a customer service representative could run an analysis to identify patterns in customer complaints and take appropriate action, or a procurement officer could assess exposures and identify alternatives following a "know your customer" alert about a supplier.

Risk modeling is another area benefiting from GenAI's democratizing influence. Quantitative modelers are already utilizing AI tools like ChatGPT to translate code between languages or to swiftly identify and link existing subroutines to complete a specific task. This makes analytical tasks more efficient and accessible.

Enhancing Communication

GenAI tools can also enable risk professionals to communicate more effectively. Despite their focus on quantitative modeling, risk professionals need strong communication to ensure their messages aren't overlooked. Indeed, format and word choice can mean the difference between emails, model methodologies and notices that are ignored and those that have a meaningful impact on mitigating risk exposures.

Sales and marketing teams are already leveraging GenAI to create more effective campaigns and to select language that resonates well with customers. It can have similar applications in risk management.

While the ability to have a GenAI algorithm independently author an email message or social media post may still be years away, today’s tools can already assist with basic editing tasks or offer suggestions for making communications clearer and more succinct. Particularly for multinational organizations or those employing non-native English speakers, GenAI can help navigate language barriers and maintain a consistent communication tone.

Parting Thoughts

GenAI tools, in short, can be used to distill information, communicate more effectively and improve the speed and quality of risk measurement and monitoring.

When financial institutions employ this technology for such tasks, they can not only streamline risk management but also gain first-hand knowledge of GenAI's benefits and potential perils. Risk teams can use this invaluable insight to strategically prioritize potential threats posed by GenAI, such as the exposure of confidential data or the propagation of incorrect conclusions because of AI hallucinations.

Using GenAI to mitigate the hazards posed by AI itself may seem paradoxical. However, just as one must secure his or her own oxygen mask before assisting others when following emergency airplane protocols, risk professionals must harness the power of GenAI to optimize their own operations before guiding their colleagues.

There are, of course, limitations to GenAI's applicability. For instance, in the United States, an AI-based credit scoring system is unlikely to be implemented soon – thanks in part to the cautious approach of regulators. The Consumer Protection Financial Bureau, for example, recently issued a directive reminding lenders about upholding fair lending standards, including providing notices of adverse action, when using AI and alternative data.

Moreover, there is trepidation among some risk professionals (particularly early-career practitioners) that the widespread adoption of GenAI tools will soon render them obsolete. But they need not worry.

While the use of “RiskBots” will undoubtedly spread, human risk managers will continue to be in high demand to complement them. Bots may be superior to humans at answering questions or processing large volumes of data, but they require human intervention to pose the right questions and guide them.

Risk managers, moreover, will also continue to hold an edge over GenAI in understanding irrational human behavior, negotiating the evolving regulatory landscape, and forecasting emerging risks.

As both external and internal company information becomes more accessible to a wider audience, the role of risk professionals will evolve. Freed of data collection and reporting tasks (with the help of AI), risk teams will be able to focus on analyzing and prioritizing emerging risks rather than simply reacting to threats.

Cristian deRitis

A greater emphasis will be placed on organizational design, including the optimization of delegations of authority that can be swiftly acted upon, eliminating the need for time-consuming and costly escalations. GenAI can lend a hand by providing frontline risk managers with direct answers to most queries while highlighting patterns and anomalies that warrant the attention of second-line risk teams.

Great success will come to risk professionals who embrace GenAI to improve their own efficiency while simultaneously striving to reduce the risks posed to their organizations by the widespread adoption of this disruptive technology.

Cristian deRitis is the Deputy Chief Economist at Moody's Analytics. As the head of model research and development, he specializes in the analysis of current and future economic conditions, consumer credit markets and housing. Before joining Moody's Analytics, he worked for Fannie Mae. In addition to his published research, Cristian is named on two U.S. patents for credit modeling techniques. Cristian is also a cohost on the popular Inside Economics Podcast. He can be reached at cristian.deritis@moodys.com.