Navigating Credit Union AI Adoption: Policies, Compliance, and Fintech Solutions
Artificial intelligence (AI) has rapidly become a fixture in our daily lives, presenting both opportunities and challenges. Credit unions need to be working now to create guardrails and policies for AI utilization. That is why the forward-thinking credit unions of Curql Collective have convened monthly since July 2023 to explore collaborative ventures in generative AI. What initially started as discourse on credit union policy has since evolved into a comprehensive analysis of case studies involving fintech and credit unions utilizing generative AI.
Importance of Policy and Guardrails
Through extensive research, the Curql Collective has delved into AI optimization within credit unions. Developing an acceptable use policy stands as a top priority in the journey of AI adoption. Equally important is providing comprehensive training to employees, emphasizing the capabilities of AI alongside its limitations. Therefore, credit unions must ensure that AI usage is closely monitored, approved, and compliant with regulations, while also avoiding unfair discrimination based on certain data.
While definitive regulations may be some time away, the credit union industry cannot afford to delay implementing robust policies for data governance and AI usage. Shockingly, less than 10% of financial institutions have a data governance policy in place. Curql strongly urges credit unions to proactively craft an acceptable use policy and conduct a thorough inventory of AI models used in their partnerships.
Prompt engineering is a strategy to safeguard AI accuracy and reliability. This process involves meticulously designing and refining inputs to AI models to yield the most valuable and precise responses. BlastPoint, a Curql portfolio company, equips credit unions with the capabilities to provide personalized member experiences at scale. BlastPoint’s expertise in AI underscores the importance of context as a critical guardrail in AI performance.
Due Diligence Considerations for AI Policy
Curql limited partner credit unions have done exemplary work around AI policy. Many keep a list of approved AI tools, and their policies advise refraining from using AI blindly for decision-making, hiring, and creating content or relying on it for important inquiries without verifying the response or getting approval. Some also have policies on storage, retention, and sharing of sensitive information in AI.
One of the primary concerns with AI is around data authenticity and security. AI needs data to perform, but that data must be high quality and from authentic sources. As AI becomes more independent, it could also reveal private information. Credit unions are rightfully concerned about how employees may store, handle, and process data used with AI. Credit unions are also concerned about the potential for AI to make errors and even discriminate.
Fintech Advantage
With a diverse portfolio of innovative companies, Curql is at the forefront of leveraging generative AI to enhance the capabilities of credit unions. Through strategic investment that shapes fintech for credit unions, Curql is driving innovation and empowering credit unions to stay ahead in an ever-evolving landscape. These advancements are particularly notable in addressing emerging challenges that pose significant risks for financial institutions.
Deepfakes are one outcome of generative AI that can be used for many good purposes but, unfortunately, also for financial crime. Illuma is addressing this issue head-on with its flagship voice biometrics product, Illuma Shield. Their approach is adaptive and dynamic and allows for risk detection.
For credit unions, AI helps us understand trends, such as patterns that can indicate fraud. Curql portfolio company DefenseStorm has a technology called Pattern Scout that identifies anomalous incidents. The technology shows analysts similar incidents across their community of partners to help with effective resolutions. Other patterns can be identified that provide valuable knowledge into member wallet share data. Facteus helps credit unions take large data sets around member transactions and leverages AI with synthetic data to gain insight into member preferences and behaviors.
AI can help members seek out self-service opportunities and find relevant answers at lightning speed. For example, Posh’s Knowledge Assistant can narrow in on a single answer that provides the member with the information they seek versus a traditional search engine that would leave the member navigating through a plethora of results. And Silvur offers a tool that allows members to get personalized answers about their retirement journey.
What's Next
As credit unions continue to explore the use of AI and apply it to more tasks, defining goals, measurements, and key performance indicators will be increasingly important. Credit unions should verify how AI fintech partners, even vetted fintech partners at Curql, use data, auditing them to ensure they are good stewards of credit union data.
At a recent conference, NCUA Chairman Todd Harper alluded to new rules for AI coming this year around fairness, consumer protection, and transparency. Until then, credit unions should continue to move forward with creating AI policies to mitigate risk. The good news is credit unions are not alone in this effort. Curql and other industry partners have compiled various resources such as proposals and recommendations on a variety of AI-related topics. At Curql we are committed to helping our credit unions navigate this exciting and expanding technological innovation through collaboration with our fintech and other like-minded industry experts.