Banks face unprecedented opportunities and challenges in today’s rapidly evolving technological landscape, particularly with the advent of generative artificial intelligence (AI).
The State of AI Today
AI has been a staple in the banking industry for decades, assisting in loan decision-making, fraud detection, marketing strategies and data analytics. However, generative AI solutions, particularly large language models (LLMs) like ChatGPT, are a more recent phenomenon. ChatGPT was released in November 2022.
Despite their initial novelty, these solutions are making substantial impacts. Economists have predicted trillions of dollars of gross domestic product growth driven by AI. The market value of Nvidia Corp. — a leading supplier of AI chips — has catapulted past that of Amazon.com and Alphabet, the parent company of Google. Vendors including Adobe, ServiceNow, Stripe, Salesforce, Abrigo, Alteryx, UiPath, UKG and NewGen have scrambled to incorporate generative AI capabilities into their products. At this point, it’s almost harder to find a vendor that doesn’t use generative AI. This disruption is real.
AI is impacting everyone, and the pace of progress and number of use cases are accelerating. Take Klarna for example: the buy now, pay later provider employs generative AI to perform the tasks of 700 contact center agents, contributing a projected $40 million annually to their bottom line. SRM has clients who are already piloting or utilizing generative AI in production environments.
The Challenges of Uncontrolled AI Use
Generative AI is also increasingly pervasive. Firms are introducing new models and tools weekly, and LLMs are increasingly integrated into ubiquitous platforms like social media, messaging apps, smartphone operating systems, browsers and applications of all kinds.
Preventing access to all generative AI solutions is impossible, given that individuals are using generative AI tools regardless of restrictions and bans. Financial institutions are adopting generative AI whether they intend to or not. This poses several risks, primarily related to data breaches and misuse of AI.
If you’re unsure where to begin, we recommend a few building blocks that will help you accelerate your plans and ensure the safe adoption of AI.
• Tailored AI policy: Developing a nuanced AI policy that addresses data privacy, regulatory compliance and risk tolerance is critical to keeping data safe and preventing unsafe use of AI. While it might seem fitting to use ChatGPT to draft an AI policy and it can act as a starting point, our experience shows that the technology falls well short of creating policies that are adequate to keep up with the shifting regulatory landscape or to provide a tailored blueprint for AI governance and adoption.
• Implement an internal LLM assistant: The most effective approach to eliminating employee circumvention of AI restrictions is to provide internal tools for staff to use that are safe with regard to proprietary data. There are various options and approaches that address a range of risk appetites and budgets.
• Employee guidance and governance: Providing guidance and education for employees, governance structures to integrate AI safely and extensible frameworks for third-party risk management and vendor selection is critical and often overlooked. AI tools are already deeply embedded in the business environment, and employees will need to learn how to use them safely. Generative AI competency will very quickly become as foundational and essential as Microsoft Excel skills.
• Vendor risk assessments: Many banks unknowingly adopt AI through their vendors. For instance, tools like Adobe Acrobat and ServiceNow incorporate generative AI features that may introduce new risks. Understanding your vendors’ use of AI helps ensure your data across various contracts.
• AI opportunity assessments: There’s immense value in identifying both quick wins and long-term value opportunities with various AI technologies, including but not limited to LLMs. Other types of AI technologies that frequently provide value for financial institutions include robotic process automation, conversational AI, biometric authentication and loan decisioning engines.
The future of banking is intertwined with advancements in AI. By taking a proactive, structured approach, your institution can realize the total value of generative AI while maintaining security and compliance.
Previous