Digital transformation and the related scope and velocity of change pose new challenges for both traditional banking and governance: from taking advantage of the opportunities related to artificial intelligence (AI), machine learning (ML), deep learning and digital platforms to managing risk, cyber security threats and relationship management. The pace and scope at which digital innovation and diffusion is taking place is far faster than financial regulators can keep up with. Further, these forces are becoming more and more integrated. According to the World Economic Forum, developments in previously disjointed fields such as artificial intelligence and machine learning, robotics, nanotechnology, 3D genetics and biotechnology are all building on and amplifying one another. Its recent survey found that 87% of companies say digital transformation will disrupt their industry but only ½ say they are prepared. It’s not surprising that some warn that “the digital realm is overtaking and redefining everything familiar”.
And in banking, the very core of our financial system and integral to our personal growth, it’s difficult to identify a banking function or line of business that does not have multiple needs for predictive analytics. And, the amount of data required for risk analysis in money lending is one of the most compelling applications of AI techniques. What’s more, it requires accurate information on the requester’s personal data, current financial situation, and credit history – alongside a loan officer’s expertise and experience in a particular community or industry. At the same time, securing funds and personal data is a prerequisite for customers. A bank’s reputation depends on their perception of its trustworthiness.
The representation, connectivity, and aggregation underlying digital transformation have introduced new ethical, legal and governance challenges. For example, ML classification tasks such as auto-acceptance of a loan or pattern recognition of applicants, such as facial recognition, can be a “black-box” for users who do not understand the features used by a particular AI model to arrive at a banking decision. Underwriting criteria produced by AI models may be too opaque, making it difficult to understand which factors drive the decision-making process.
Yet, while AI and ML innovations pose both a risk and an opportunity to banking as we know it today, 61% of the directors who responded to a recent National Association of Corporate Directors (NACD) survey report they would be willing to compromise on cybersecurity to achieve business objectives, while only 28% prioritize cybersecurity above all else. Additionally, cybersecurity experience was cited in only 2% of newly appointed directors.
As AI and ML in banking increases customer awareness of data ownership and security, banks should consider creating a governance model that ensures participating third parties cannot damage their reputation. The board’s role in cybersecurity oversight is also evolving and is more important than ever, especially as ransomware attacks and other security breaches rise in frequency and in material impact. Finally, the board needs to focus first on the quality of the reporting it receives from management about risk.
In my book, I have a tabletop exercise about this very issue – digital transformation and banking.