Five Banking Trends Driving Artificial Intelligence Forward

Governance of AI has been found to be more effective if it embraces diversity of skills and perspectives and covers the full spectrum of functions. This cross-functional mitigates the AI complexity and related data challenges. 

Owing to the nature of the XAI algorithm and the fact that outcomes and drivers of XAI are fully explainable, the uncovering and understanding of latent data quality issues can happen during XAI model developments, which supplements up-stream data management efforts creating a virtuous circle.

AI models can create risks that arise from complex inputs, inter-variables relationships, the algorithm itself, and even the ensuing outputs. Hence, explaining AI model output is critical and more so in the financial industry as regulators keep a keen watch to proactively prevent the pervasive use of AI from causing a systemic risk to the banking system.

While AI can be deployed in many business scenarios, and that has been the case, the barriers to AI adoptions, especially for smaller banks, can still be significant. High on the list are the barriers involving data, documentation, explainability and governance. AI can magnify existing risks and introduce new challenges resulting in more controls and structures.

“Explanation” depends on context, a good explanation for a customer is very different from one to data scientist or bank executive. This explainability gap require consistency to facilitate understanding among the different stakeholders and there is an already increasing calls for this level of explainability.