Bill Gates said in a blog post: “The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. Entire industries will reorient around it. Businesses will distinguish themselves by how well they use it.”[1]

Trends

The evolution of AI within the banking sector is progressing from exploration to comprehensive integration. To unlock the transformative potential of emerging artificial intelligence and machine learning advancements, banks are urged to transcend mere speculation and embrace the tangible applications of AI. Banks need to explore the practical avenues for implementing AI and learn the strategies for ensuring successful execution.

Banking regulators are closely watching the evolution of AI and its relationship to banking. While there are no banking regulations specifically targeted at AI yet, the Consumer Financial Protection Bureau (CFPB) issued guidance last September regarding banks that utilize AI and related complex modelling in their consumer credit decision-making processes.[2] The CFPB guidance warns that precautions are necessary to protect consumer access to nondiscriminatory credit decisions given the continued surge in AI use for consumer credit decisions—and its concomitant algorithms, machine learning and voluminous data processing. The CFPB made clear that banks must be able to specifically explain their reasons for denial. The decision must be tailored to the action and there is no special exemption for artificial intelligence. The CFPB's guidance further reiterated banks must comply with the Equal Credit Opportunity Act and Regulation B by providing accurate and specific explanations when they take adverse actions against consumers, regardless of AI use that may, by nature, be difficult to explain.  While this guidance focuses on adverse action notices, it also highlights the importance of using AI as a decision-making tool with proper checks and balances, rather than relying on it explicitly.  As pointed out in the recently released report on AI from the US Department of Treasury[3], like all emerging technologies, the use of artificial intelligence poses risks to the banking industry. The wise bank management team will be the one that moves into the realm of AI armed with sound policies, procedure and processes and well-trained employees, ready to identify and mitigate the risks posed by this exciting technology.

Use Cases

Assessing the benefits and drawbacks of adopting a more extensive enterprise AI strategy is step one in that expansion and execution. As banks contemplate making moves in this space, delving into practical use cases that our team has been exposed to can provide invaluable insights for decision-making. By examining current use models, executives can make well-informed decisions customized to their banking institution’s unique requirements, thus maximizing any capital expenditures. Use cases do not cover every possibility, they simply demonstrate a variety of ways in which AI is currently being used enhancing value across a banking institution.

1. Customer Experience
   • Data-driven AI, which can allow micro-segmenting of existing customers and prospective customers, enables banks to better predict preferences and behaviors, providing tailored solutions that foster deeper engagement, opportunities for cross-selling and avenues for innovation.
2. Service
   • Advancements like conversational bots handling basic inquiries or "smile-to-pay" identification for seamless transactions are enhancing customer service. Bots can also harness historical data to provide personalized discussions, including social media, and provide operational efficiencies using call center patterns and calculating to the second customer wait times. Additionally on the service side, SSO (single sign-on) applications using AI enhanced authenticators are being used to improve security protocols.
3. Collections
   • AI has the potential to boost efficiencies and develop proactive strategies to assist both customers and lenders. Banks can capitalize on customer data to detect early warning signs of potential delinquencies and defaults, anticipate reasons why customers may fall behind on payments and provide tailored solutions to prevent such occurrences.
4. Underwriting
   • Combining robotic process automation with machine learning models and a range of data sources can speed up the loan underwriting process and enhance risk evaluation. Automating tasks such as document scanning and manual data collection streamlines the process of gathering pertinent information. Machine learning models can then analyze data from multiple sources to precisely assess borrowers' risk profiles, enabling quicker loan decisions.
5. Compliance
   • Banks can enhance efficiencies and reduce costs by harnessing AI to automate labor-intensive procedures and swiftly identify changes to the regulatory landscape, ensuring continual compliance. Further, machine learning models may spot consumer compliance issues, trends and patterns unseen by even the keenest human compliance review team.
6. Risk
   • Fraud detection stands as the most utilized application of AI in banking today. Banks are experiencing the advantages of such applications, not only through decreased losses and more streamlined resource allocation, but also in terms of customer experience. For example, credit card companies utilize transaction and authorization data to enhance the precision and speed of fraud prediction and detection. By minimizing false positives, fewer legitimate transactions are halted, thereby enhancing the overall customer experience.

AI Readiness

Market research entity Avanade reports that 92% of organizations will need to shift to an AI-first operating model by the end of 2024 to keep pace with competitors. [4] Further, the IAPP-EY Privacy Governance Report of 2023 shows AI rocketing to the top of list for most important strategic priorities in privacy for the future. [5]

To stay abreast of AI marketplace trends and confidently navigate the future, banks should prioritize several key actions:

1. Embrace Data-Driven Innovation: Recognize the potential of AI technology to drive innovation at an accelerated pace. Utilize AI to enhance operational efficiency, facilitate growth initiatives, differentiate services, address risk and compliance requirements and elevate customer experiences.
2. Adapt to Evolving Technology: Acknowledge the changing landscape of AI technology, including the decreasing costs and barriers to adoption. Stay informed about advancements in AI tools and methodologies, and be prepared to integrate these technologies into existing systems and processes.
3. Make Strategic Investments: Allocate resources strategically to support AI initiatives, focusing on areas such as cloud computing, big data platforms and updated data architectures. These investments can help streamline AI development, deployment and scalability without necessitating significant upfront capital expenditures.
4. Address Operational Challenges: Recognize and address operational and organizational hurdles associated with AI implementation. Invest in developing necessary skills among staff and ensure effective integration of AI technologies into the broader organizational framework.

By taking these proactive steps, banks can position themselves to leverage the full potential of AI technologies, driving innovation and maintaining competitiveness in an increasingly data-driven landscape.

Impact

As a whole, the use cases show that AI is becoming essential for business success in the banking world and beyond. While enthusiasm for AI is robust, banks are still adopting a cautious approach. This is rooted in the fact that while the technology is undeniably impressive in its capabilities, its applicability is still not flawless.  Banks want and need AI to be tested and proved before adoption.  Furthermore, banks are likely to face heightened regulatory scrutiny regarding the "explainability" of AI systems and their decision-making processes. They will need to establish protocols that enable users to comprehend the outputs generated by machine learning algorithms. Embracing explainable AI as part of a responsible AI implementation strategy offers increased transparency, enabling the identification and rectification of potential flaws and vulnerabilities in models. This approach not only enhances the performance and accuracy of AI models but also ensures fairness and transparency. Banks that actively engage in developing AI algorithms with robust explanatory capabilities will be better positioned to earn the trust of both customers and regulators.

At the end of the day AI is simply data, and the key to all data is governance – how this data is collected, stored and used is critical. Without the proper understanding, organizations incur risks, including reputational risks, data risks and security risks.

For questions regarding this evolving landscape, or if you require assistance updating your banking cybersecurity policies to account for AI risks, contact Dinsmore attorneys Christian Gonzalez, Craig Horbus or Michael Dailey for more details.