Key Considerations for a Smarter, Safer AI Strategy

Connor Heaton, Director of Artificial Intelligence, SRM

Artificial intelligence (AI) is creating a seismic shift in the financial services ecosystem. Typically, use cases for AI were limited to specific tasks like loan decisioning, fraud detection, or sentiment analysis. However, the rise of generative AI and Large Language Models (LLM) has democratized the field, making AI tools user-friendly, cost-effective, and applicable across various domains.

Widespread AI adoption is inevitable, particularly as high-profile organizations begin to leverage generative AI capabilities. For instance, payments giant Stripe successfully leveraged ChatGPT-4 to automate monitory capabilities, flagging fraudulent activities within community and support forum posts. Additionally, Morgan Stanley has used generative AI to handle federated search capabilities and the company's internal knowledge base. 

The Credit Union Connection has partnered with @igniteFI to host a unique webinar series, The Fintech Adventure! Our first virtual event will tackle a deep dive into AI for an informational hour you can’t miss on Sept. 19. Then, on Sept. 20, follow it up with speed dating sessions with the Credit Union Business Partners in the trenches. Space is limited, so learn more and register for Day 1 here and Day 2 here!

It has been standard practice for credit unions to leverage third-party AI systems to manage a handful of tasks. However, generative AI technologies' widespread availability and versatility require credit unions to implement policies and tools to oversee AI systems that perform several diverse functions.

Adopting AI brings common challenges

Around 75% of employees secretly use AI tools in the workplace. Employees may find that these tools help streamline and enhance their productivity, but unchecked usage can place credit unions' data at unnecessary risk. Generative AI is pervasive, having been weaved into apps, web browsers, and ubiquitous platforms like Microsoft Office, making it crucial to set guardrails entailing appropriate use at a scale. 

Mitigating risks around data leakages is another major challenge for credit unions. Although generative AI tools like ChatGPT and Copilot are easily accessible and free, they also harbor inherent risks. Employees or vendors can unknowingly expose data by entering sensitive information into these platforms. 

Generative AI has also made fraud quicker and easier for criminals. Credit unions have effectively combated fraud for decades, but their members are less experienced and, therefore, more susceptible. To help protect their members, credit unions are investing in resources to reduce member vulnerability to typical fraudulent schemes. 

Vendors like NewGen, Adobe, and Alteryx have also added generative AI features to their products. Compliance issues are bound to arise without proper protective clauses in existing contracts. Being cognizant of these additions lets credit unions know how their data is used across multiple contracts.

Credit unions hesitant about AI implementation should consider building a solid foundation for adoption. This is accomplished by creating a comprehensive adoption policy addressing risk management, data privacy, and regulatory compliance to prevent internal misuse and data leakage. Taking a step further, credit unions can also deploy an internal LLM assistant via a secure vendor for proprietary data, such as Azure, OpenAI, or Microsoft Copilot.

Simplifying the vendor selection process

As vendors scramble to integrate generative AI technologies into their offerings, the market has become flooded with various LLM-based products. Credit unions are struggling to keep pace, and without proper expertise or guidance, institutions can easily become overwhelmed by choice during the vendor selection process.

Understanding the construction of available AI solutions on the market allows credit unions to filter through thousands of vendors. Most solutions available to credit unions are built on a select few foundational models, with vendors only making incremental changes. At their core, these solutions share over 80% of their DNA with foundational models. This knowledge narrows the vendor space and, from a risk perspective, makes the vendor selection process far more manageable. 

AI will continue to disrupt the credit union space as more institutions implement generative AI solutions. Now more than ever, it is mission-critical to implement a structured and proactive AI strategy that prioritizes security, compliance, and careful vendor selection.

Connor Heaton is the Director of Artificial Intelligence at SRM, an advisory firm serving financial institutions in North America and across the globe. He leads client engagements focused on artificial intelligence and helps organizations understand and adopt disruptive technologies.