Banking in the AI era: The risk management of AI and with AI

Artificial intelligence (AI) is instrumental in helping banks gain a competitive edge with personalized client engagement while building operational efficiencies. However, innovating with AI-powered tools must be balanced with a prudent approach that meets rigorous regulatory requirements. A reflection on risk and compliance practices in the era of AI becomes a prerequisite to scale the power of AI across financial institutions.

A new IBM Institute for Business Value (IBM IBV) survey of risk, compliance, and validation (RCV) officers reveals a telling snapshot of where these bank executives see AI’s greatest promise—and its most daunting challenges. The findings are as illuminating as they are provocative, with 61% of respondents zeroing in on fraud risk detection as the area to harvest the most significant boost to business value within their functions.

The ability to detect and prevent illicit activities in real time is nothing short of mission critical. Close behind, 52% of survey respondents highlight cybersecurity, another domain where the margin for error is razor thin.

61% of bank executives say fraud risk detection will provide the biggest boost to business value, with cybersecurity close behind at 52%.

The toughest transformation tasks

Fraud detection and cybersecurity are top of mind when thinking about the benefits of AI, but processes related to know your customer (KYC) and anti-money laundering functions (AML) are especially thorny challenges. When asked about the complexity of using AI systems for key RCV processes, 43% of executives identify KYC and AML as the most daunting tasks to transform with AI tools.

KYC and AML require navigating a maze of global regulatory requirements and verifying vast amounts of data with pinpoint accuracy. It’s not just about crunching numbers; it’s about understanding context, interpreting behavior, and making judgment calls—all in the shortest time possible.

This is where agentic AI becomes a game changer. Unlike machine learning and generative AI, which follow more specific task definitions, agentic AI autonomously orchestrates sub-agents that learn, optimize, and adapt on the fly. Agentic AI holds the potential to not just flag potential risks and suspicious transactions but actively investigate financial crime as it cross-references datasets and analyzes patterns. The potential is tantalizing: a world where compliance is not just a box to check but a dynamic, proactive shield against financial crime and a constantly updated tool for regulatory compliance.

Compliance documentation is the perfect proving ground where banks can build confidence in their AI capabilities and gather data needed to scale up to more complex applications.

Creating proving grounds for scaling AI

RCV officers report that investments in documenting compliance are among the most prioritized, even though they rank lower in terms of expected business value. Although this seems counterintuitive, compliance documentation aligns with the practice of risk management—managing the risks of banking along with the risks of innovation.

Surveyed banking executives indicated which initiatives and use cases they believe will empower RCV functions to improve the scaling of AI technologies across their institutions, while managing associated risks. These responses underscore three pivotal initiatives:

• Stress test simulations. Selected by 63% of respondents as the foremost priority, simulations are vital to rigorously assess AI-powered system viability and reliability. They proactively identify potential weaknesses in AI models before operational deployment.
• Real-time risk controls. Endorsed by 48% of respondents, these mechanisms provide ongoing risk assessment and monitoring. They also have capacity to take immediate corrective measures to help ensure AI systems don’t drift and hallucinate.
• Talent and skill development. Noted by 33% of respondents, this emphasizes the necessity of fostering workforce proficiency to oversee and enhance AI-driven technologies.

Download the report to help your banking industry clients involved in risk, compliance, and validation functions to scale AI across their financial institutions, while managing the associated risks.