Internal audit and AI: how human collaboration meets innovation

Recent developments in – and media coverage of – advances in artificial intelligence, especially large language models (LLMs), require internal auditors to adapt urgently.

Yet anyone who has ever tried to acquire a new skill or develop a new habit knows that change is difficult. Improvement in any area requires commitment, focus, and discipline. However, it seems harder than ever to harness these traits in our modern age of instant gratification, information overload, and constant distraction. Technology appears to be both the problem and the solution – but where is the human element?

In this brief article, I will discuss what I see as all-too-human barriers to progress and suggest ways to overcome them. In September, I will speak at an AuditBoard webinar on this topic—a great opportunity for readers to discuss it further.

Progress ... and pitfalls

We all know that technology works for good or ill and that its development and use depend on us—the humans! We know the old software development saying, “Garbage in, garbage out,” and this applies to audit management software as well as LLMs. We cannot blame software for producing poor-quality results if we haven’t first used critical thinking to refine our analysis and then plain language to articulate it.

But it’s hard. We are short on time, energy, and concentration spans. Spending precious moments thinking through our work and checking the output seems a waste of time. Why do the work yourself when there are lovely LLMs available to produce text from some hastily scribbled prompts?

I will leave aside for now the question of information security and false economies, although we will cover these points in the September webinar. What I want to raise here is a difficulty I haven’t seen mentioned elsewhere: the fact that we are asking people to use a skill they have long neglected or possibly never acquired.

I am not a digital native. I first used a computer as an undergraduate in the late 1980s, around the time I learned basic coding. My first email address dates to—I think—soon after, but I don’t recall using it. Research took place in libraries, using physical card catalogs and then databases.

This means that I learned to read, research, reason, and write using age-old technologies. It was slower, but it developed certain mental abilities, such as memorization, and encouraged curiosity. Previously, it was commonplace to stumble across books and journals you hadn’t known of while flicking through a card catalog or wandering through library stacks, browsing the shelves. Later, internet-based research streamlined searches, but at the expense of serendipitous discoveries.

Nowadays, I, like almost everyone else, depend on the internet and my smartphone for virtually everything. Switching off to focus on a task is rare, and the last physical card catalog I saw was a museum piece!

We need to maximize our use of technology by applying the “old” skills of rigorous research, critical thinking, and plain language. Yet this is triply difficult: those of us who learned these skills have lost many of them, letting our mental muscles atrophy. Those who didn’t learn them may struggle to acquire them. And all of us may see doing so as a waste of time when “tech” can do it all for us.

We live in a technogenic environment, which is like living in an obesogenic environment that encourages fast, easy food options at the expense of our well-being. We have the wonders of technology—whether a smartphone, search engine, or LLM—at our fingertips, and so they become the default. But are we using them thoughtfully, judiciously, and critically? It seems not, with daily accounts and even studies showing humans falling prey to misinformation and including LLM hallucinations in professional documents.

Moving forward together

I’m not suggesting that modern technology is junk food – far from it. However, just as we should think carefully about what we put in our bodies, we should do so with what we put in – and extract from – our minds.

This is a tremendous opportunity for all of us, as a society and as internal auditors. By using the “old” skills mentioned above, we can create better prompts and outputs and improve future iterations and even software releases.

We can also use what some might see as generational differences to share knowledge and skills, to better collaborate on all aspects of our work. If – to use crude but popular terms – Gen X and Gen Z pool their respective strengths and skills, the entire team and organization benefit. It seems obvious, but why wouldn’t you have an “analog” native explain to a digital native how using paper and pen to, for instance, outline or mind-map a topic, before trying the same exercise using software and seeing the differences? Similarly, digital natives can show non-digitals functionalities and approaches we never suspected.

Many organizations use reverse mentoring, allowing junior or younger staff members to share their perspectives and insights with more senior colleagues. Pairing internal auditors with complementary skills and mindsets is a common way of conducting engagement interviews and testing. What one colleague may overlook or take for granted, the other can question or use as a development point. And of course many internal audit functions of all sizes have regular informal training sessions: team meetings, lunch-and-learns, even shared drives where colleagues post useful resources or successful communications.

Working together to reason rigorously and articulate clearly, rather than outsourcing thought and language to software, means we control the narrative, even under pressure. Keeping abreast of developments in technology with an open yet critical mind means we can benefit from the best aspects while being alert to and mitigating gaps or weaknesses.

If we start seeing “old” and “new” skills as necessary and complementary rather than competing, it can only improve everyone’s abilities, morale, and software skills.

So, what’s next?

Register for my 23 September webinar with AuditBoard, Internal audit communications: Authentic intelligence and artificial intelligence. Look out for future articles from me with AuditBoard. Topics will include critical thinking and plain language, and the Institute of Internal Auditors’ new Global Practice Guide on final engagement communications.

About the authors

Sara I. James, PhD, CIA, is the owner of Getting Words to Work®. With over 35 years’ teaching, writing, publishing and corporate experience in the US and Europe, Sara provides report-writing training worldwide. She has written numerous articles on language and reporting, and spoken at national and international conferences. As a member of the Chartered Institute of Internal Auditors’ Technical Guidance Working Group, Sara provides resources for members in the UK and abroad. Connect with Sara on LinkedIn.