The artificial intelligence (AI) market is growing rapidly at more than 42% annually, with recent MIT research indicating over 90% of larger enterprises are using AI to improve customer interactions. This growth and adoption will, in turn, lower cost to entry, putting AI-empowered solutions and products within the reach of many credit unions.
Companies at the forefront of AI are coming to the realization that, as with many new technologies, AI introduces new risks to their organizations. Accordingly, these companies are increasingly including disclosures surrounding AI and its potential impact in both their regulatory filings and their shareholder communications.
Google parent company Alphabet noted in its annual financial filings over the past few years that investments in new technologies are inherently risky and may disrupt current operations or harm financial conditions and results. Microsoft has included similar statements in its filings, divulging the risks inherent in the deployment of AI technologies. While these references might seem insignificant within the myriad of other disclosures made in regulatory and financial filings by large public companies, they show that the adoption of new technologies like AI presents new and unique risks that might not have previously warranted mention.
As the adoption of AI-empowered solutions and products continues to expand across the credit union industry, financial institutions will be opening themselves up to these new risks. Though not of the magnitude faced by companies like Google or Microsoft, many financial institutions need to reexamine existing policies, procedures and practices in order to ensure compliance and mitigation of additional risk exposure. Combining new technologies with the expectations of recent federal and state reporting requirements surrounding its use, and the implications on credit unions, their leadership and their boards of directors become more prevalent.
What can credit union boards do to acknowledge and address the new risks and added responsibilities driven by AI adoption?
1. Understand New Technologies
Familiarization with new technologies and how they can potentially be deployed throughout the credit union space is vital. Most consumers know AI exists and may realize they interact with it on a daily basis, but they would not necessarily consider themselves data literate or AI fluent. Schedule a presentation from an AI expert at the next board retreat or planning session, as well as curate educational materials that can help establish foundational understanding. Knowledge of basic terminology and concepts will prepare all board members for the oversight for which they will be responsible.
2. Evaluate Existing Processes
Some credit unions may already be addressing the new risks associated with the use of AI. Many institutions have a well-defined risk management or enterprise risk management team that, in their normal course of business, would have recognized and begun working to document, qualify and quantify potential impacts. Some might have already brought forward recommendations for leadership or board action to address AI. Board members can look across the organization to identify what actions the credit union has already taken to better position itself in the light of these new risks. Consider these questions: Do audit reports reference the existence of AI-empowered solutions or products? Has the board recently reviewed an enterprise data strategy or been asked to approve the hiring of a chief data officer? Has the credit union documented an AI framework of best practices, created an AI governance council or established an AI center of excellence?
3. Pose More Questions
Beyond organizational readiness, board members should become familiar with other aspects of a credit union’s use of AI. Gaining insight into cases of AI use across the organization is a great start. Where is the credit union using AI, and what problems is it intended to solve? Awareness is critical, but deeper understanding is also important. Who owns the AI models being used? What specific risks might be associated with a particular model? What type of training has taken place? What steps have been implemented to ensure data privacy and avoid model bias?
Ultimately, just as the board has the responsibility to know a credit union’s loan portfolio is sound, it should also have a vested interest in the credit union’s AI portfolio. A well-established and well-administered enterprise AI program will be able to provide key performance indicators, including return on investment, models in production, efficiency and efficacy gains, and projection or prediction accuracy rates.
Credit unions are on a collision course with a subject with which they are not intimately familiar. Among the many advantages AI can bring to a credit union come a bounty of new responsibilities associated with its inherent risks, and board members have a duty to start preparing themselves and their organizations for these unique challenges.