At Possible 2023, leading artificial intelligence (AI) researcher, writer, and thinker Avi Goldfarb riffed on psychologist Daniel Kahneman’s now-famous quote: “Clearly AI is going to win. How people are going to adjust is a fascinating problem.” Goldfarb quipped that for those of us who had not already won a Nobel Prize, theproblem was not fascinating, but terrifying.
Extending this thought to the sector I’m most familiar with, I’d agree: Most banks are terrified of the ramifications of AI for their business and their sector. This is not the AI and machine learning (ML) with which many have become familiar over the past decade. Banks are among the leaders in deployment of rules-based algorithms to make decisions on everything from next best offers to anti-fraud interventions. The emergence of self-learning generative AI has the potential to upend and transform almost every aspect of the banking sector.
Transformative potential
Drawing an analogy with the electrification of factories in the early 20th century, Goldfarb showed that those who merely replaced a steam engine with an electric motor saw limited benefits. They saved some money from lower energy costs, but the electric motor did the same job as the steam engine—just more efficiently. This is the stage at which many banks find themselves today: Most generative AI deployments are intended to reduce cost or increase speed of existing processes.
The real transformation from electrification came when businesses recognized that electrical power had the potential to completely reengineer work. By the 1920s, electrification was freeing factory pioneers to dramatically improve production. This system-level change created the flexibility to adapt, innovate, and grow—delivering enormous value from investments in electrification. We are at that inflection point with generative AI now—and unlike electricity, it won’t take 40 years to transform financial services.
From point solutions to system-level change
Banks must ensure they aren’t left experimenting with generative AI point solutions while others reinvent the sector. But, as a heavily regulated industry, how should banks embark on this transformation? How can they align governance and regulatory reporting requirements with “black box” generative AI solutions? And, with the fast pace of technological change and regulatory responses still in their infancy, what can be done now to prepare for a highly uncertain future?
Good data diet
Data sustains generative AI—you may not be able to predict how a self-learning algorithm will develop, but you can control what you feed it. Banks should follow three vital steps to prepare for the AI-driven future:
Build
Build and maintain enterprise data platforms that provide unified, trusted data sources
As generative AI develops and consumes source data, data platforms must be capable of ingesting, integrating, verifying, and managing data from diverse external sources—not just a bank’s own data. Provide stringent governance around provenance, protection, and security of data.
Rule_Settings
Establish well-defined and tested processes for managing data relevance and reliability
To understand how and why generative AI processes make specific decisions, banks must know which data was used to train a model, and which datasets were consumed to make the decision. They must also show how they monitor and manage drift in models as they learn, so “hallucinations” are quickly spotted and rectified.
Verified
Instill good data practices
Enhance data cultures to build best practices for confidently working with data and AI. Generative AI has the potential to amplify, improve, and reimagine virtually any banking role. Train staff to work effectively with AI to create real value for the bank, its staff, and its customers.
Tomorrow’s leaders will take bold steps today to redefine their businesses with generative AI. To prepare for tomorrow, banks must invest in robust, scalable, and flexible data platforms to support advanced analytics and generative AI.
Next, we’ll explore a specific example of delivering an AI-centric strategy that reimagines fundamental aspects of banking.
