Internal Audit's New RSI: Recursive Self-Improvement with AI

Beyond Automation: AI for a Learning Audit Function

The traditional understanding of RSI, Repetitive Strain Injury, highlights the pitfalls of excessive repetition without variation. In a similar vein, Tom McLeod argues that Internal Audit must now contend with a new form of RSI: Recursive Self-Improvement, driven by Artificial Intelligence. This shift moves beyond simply automating existing audit tasks to leveraging AI for continuous methodological enhancement.

Many internal audit functions currently view AI as a tool for basic automation, such as drafting audit programs, summarizing policies, or generating reports. While these applications offer efficiency gains, McLeod emphasizes that the true transformative power of AI lies in its ability to foster a "learning audit function." This means AI would not just execute tasks, but actively analyze performance, identify weaknesses, propose improvements, and integrate these learnings into future audit cycles.

Imagine an internal audit function where:

• Every audit informs and refines the next.
• Findings sharpen subsequent risk assessments.
• Weak management responses lead to improved challenge questions.
• Control failures strengthen the testing library.
• Audit committee feedback enhances reporting.
• Incidents and regulatory changes instantaneously reshape the audit universe.

This vision transcends mere automation; it describes an audit function that evolves and adapts at an unprecedented pace. The critical question for Chief Audit Executives is not whether AI can save time, but whether it enables the audit function to learn and adapt faster than the organization's risk landscape is changing. Relying solely on AI to speed up outdated methodologies risks automating an obsolete model, which ultimately provides little strategic value.

Citation: Tom McLeod
Source: LinkedIn