MDIs in the Age of AI

All In, but First Let’s Have a Working AI Policy 

Minority Depository Institutions in the Age of AI: Adopting New Technologies

The banking landscape is constantly evolving, driven by rapid advancements in technology. For Minority Depository Institutions (MDIs), staying at the forefront of innovation is crucial to serving communities effectively and ensuring long-term growth. Recent insights gathered from members of the National Bankers Association highlight a clear trend: MDIs are actively considering and exploring the adoption of new technologies, particularly in areas like Artificial Intelligence (AI) and digital banking services.

A look into the priorities of NBA members reveals significant interest in understanding how technologies like AI can be leveraged in banking operations. When ranking various industry topics, “AI in Banking: Emerging Use Cases” was frequently placed among the top interests by NBA members. This indicates a forward-looking perspective within the MDI community regarding the potential impact of AI in the coming years.

Delving deeper into the specific applications of AI, survey respondents expressed varied but clear interests. Conversational chatbots, which can support functions like loan applications or customer service calls, emerged as a top-ranked AI use case. Other high-interest areas included fraud prevention, specifically deploying machine-learning algorithms to enhance security. Generative AI, aimed at providing individualized recommendations to customers, also garnered significant interest. This detailed view underscores that MDIs are looking at practical, customer-facing, and operational AI applications. Machine learning—which includes fraud detection and predictive analytics—is already used by core processors for services such as internet-banking fraud anomaly detection and card-payment fraud. Natural Language Processing (NLP), which can understand and generate human language, can also be utilized in fraud situations and for applications like chatbots.

Beyond AI, digital-banking services remain a critical focus for MDIs. Reducing friction and increasing conversion in digital channels is consistently ranked as a high priority. Specific areas of interest within digital banking include optimizing processes for consumer account opening and business account opening. There is also considerable interest in enhancing digital capabilities for business-loan origination and consumer-loan origination. These rankings demonstrate a commitment to improving the digital experience for both individual and business customers. Furthermore, interest in topics like data warehousing for real-time access to performance indicators and offering instant payments like FedNow and Zelle further illustrates the breadth of technology adoption being considered. Discussions around open-banking solutions for real-time deposit-account verification for consumer or SME loans also highlight interest in emerging technologies.

While the interest in adopting these new technologies is high, the path to successful implementation requires careful navigation and a strong focus on responsible and ethical use. Utilizing AI can distill information and help find solutions, saving time. However, there are significant risks for financial institutions in using this technology. These risks include uncertainty about who owns AI-created content and security/privacy concerns when inputting proprietary company information or sensitive data about employees, customers, or consumers. Additionally, the accuracy of AI-created content must be scrutinized, as it may be outdated, misleading, or even fabricated. While there are different types of AI like robotics, machine learning, computer vision, expert systems, and Natural Language Processing, many risks are particularly associated with generative AI, which can create new content.

To address these risks and ensure safe adoption, establishing clear policies and expected controls is crucial. MDI members are encouraged to form multi-layer work groups before deploying generative AI for work tasks. It’s essential to protect personally identifiable information, and no company, client, or PII of any kind may be submitted into general AI platforms. PII includes information that can identify an individual, such as names, addresses, Social Security numbers, and financial information. Robust data governance is essential for AI systems, ensuring data quality, protecting data privacy, and complying with relevant regulations like the Gramm-Leach-Bliley Act (GLBA), which prohibits the release of nonpublic personal information. PII cannot be shared without explicit consent, except as required by law.

Furthermore, all AI-generated content must be reviewed for accuracy before being relied upon for work purposes. If a reliable source cannot verify the information, it cannot be used. Training employees is also vital to help them spot AI-related fraud, identify unusual activity in AI systems, and understand reporting protocols for potential security incidents.

Acceptable uses of AI, within a controlled environment, can include researching general knowledge, brainstorming ideas, creating formulas, drafting emails or letters without entering customer PII, or finding solutions to technical problems. However, using unreviewed AI-generated text in final work products, submitting consumer content or PII into AI tools, or claiming AI output as original thought are unacceptable uses. Adhering to these guidelines is critical, as violations can result in legal liabilities.

The MDI community recognizes the importance of innovation and safe technology adoption, as highlighted by initiatives like seeking members to assist the Innovation Committee, chaired by Dom Mjartan, CEO of America Pride Bank, with program selection and vendor evaluation. This collective effort points to a need for collaborative support in assessing and adopting new technologies responsibly.

In conclusion, MDIs are not merely observing the technological changes in the banking industry; they are actively engaging with them. From exploring the potential of AI for customer service and fraud prevention to enhancing digital account opening and loan origination, the drive for innovation is clear. However, the journey involves complexities and significant risks that must be managed through careful planning, robust policies, employee training, and rigorous data handling. MDIs still need support, particularly in identifying risk-adjusted adoption use cases that align with their unique missions and operational capacities while ensuring compliance with privacy and security standards. Continued collaboration, knowledge sharing, and access to tailored resources will be essential in helping MDIs successfully navigate the age of AI and leverage technology safely and effectively to better serve their communities.

Learn how to implement a corporate-wide AI policy that ensures regulatory compliance, secures your data, and follows best practices for risk management in our full guidance document here.