I caught up with a former colleague at an internal audit conference a few weeks ago. He’s now a chief audit executive, and like most conversations these days, ours turned to AI. We talked through the usual ground: how AI will change testing, how it will reshape reporting, how it will free up hours for judgment work. Then he said something that stopped me.

“It’s always been about getting to the truth in our audits,” he told me. “But today it is harder than ever.”

I asked him to explain. He said a corrosive culture is only one obstacle among many. The bigger problem, he argued, is that truth itself has become a moving target. Society no longer agrees on a single version of events. Even news organizations shade their reporting to match the politics of their audience, so two people can watch the same event and walk away with opposite conclusions. He said people are learning to treat truth as something nuanced, something buried in noise rather than stated plainly. And he doesn’t think AI helps. AI still hallucinates, and even when it doesn’t, it pulls from sources that are themselves unreliable. Feed a model tainted information and it will hand you a confident, well-written, wrong answer every time.

I’ve thought about that conversation ever since. He’s right that our profession has always chased truth: the true state of controls, the true condition of a process, the true risk behind a decision. When I was President and CEO of The IIA, I even identified publishing an erroneous report as one of the “7 Deadly Internal Audit Sins.”

What’s changed when it comes to nailing down the truth is the ground we chase it on. Here’s my read on why truth has become such an endangered species, and what I believe internal auditors need to do about it.

5 Reasons Truth Is an Endangered Species

• Information moves faster than verification. A claim can circle the globe before anyone checks it. By the time a correction arrives, the original version has already shaped opinion. Audit cycles run on weeks and months. Rumor and misinformation run on minutes.
• Algorithms reward engagement, not accuracy. Platforms surface content that keeps people watching, and outrage keeps people watching better than accuracy does. Over time, that trains people to expect information that provokes a reaction rather than information that simply informs.
• Confirmation bias has a delivery system. People used to have to look for information that confirmed their views. Now that information looks for them. Personalization technology narrows what each person sees, and a narrower feed feels like a fuller picture than it actually is.
• AI blurs the line between synthesis and fabrication. A model can generate a citation, a statistic, or a quote that sounds authoritative and doesn’t exist. Lawyers have already been sanctioned for filing briefs built on fabricated case law a chatbot invented with total confidence. Auditors face the same trap.
• Trust in institutions has eroded. When people distrust courts, regulators, and the press, they stop looking to any shared referee for what counts as fact. Internal audit is an institution too, and it has to earn the credibility it wants to lend to the truth.

5 Ways Internal Auditors Can Ensure the Accuracy of Their Work and Reports

• Trace every material fact to a primary source. Don’t accept a summary, a dashboard, or an AI-generated draft as your evidence. Go back to the system of record, the signed document, or the original data extract. If you can’t trace a fact to its origin, treat it as unverified.
• Corroborate before you conclude. One data point is an observation. Two independent data points that agree are evidence. Build the habit of triangulating before you write a finding, and be honest with yourself about the difference between a pattern you’ve confirmed and a pattern you’ve assumed.
• Treat AI output as a draft from a junior team member. Review it the same way you’d review work from a new hire: check the sources, test the logic, and verify every number before it enters a report. A tool that writes with confidence isn’t the same as a tool that writes with accuracy.
• Document your reasoning, not just your conclusion. If a stakeholder challenges a finding six months from now, you need a clear trail showing how you got there, not just where you landed. Writing out the reasoning also exposes any gaps.
• Protect your independence from the loudest voice in the room. Corrosive culture pressures people to shade findings toward what leadership wants to hear. Your job is to report what you found, not what will play well in the meeting. That takes a kind of courage that no methodology can substitute for.

Stop, Look and Listen

My former colleague made a point that I haven’t stopped contemplating. Getting to the truth used to mean gathering enough evidence to reach a defensible conclusion. Now it also means defending the idea that a defensible conclusion is even possible in a world flooded with noise, bias, and synthetic content. That’s a heavier burden than our profession has carried before.

Mark Twain put it better than I can: “It ain’t what you don’t know that gets you in trouble. It’s what you know for sure that just ain’t so.” That line should hang the wall of every audit department in the world. The riskiest moment in any engagement isn’t when you’re missing information. It’s when you’re certain about something you never actually verified. AI makes that risk sharper, because it delivers wrong answers with the same confident tone as right ones.

I don’t think the answer is to distrust every source or every tool. AI can still make our work faster and, used carefully, more thorough. The answer is to hold our standards of evidence even higher as the volume of unreliable information grows around us. Internal audit has always been in the business of separating what’s verified from what’s merely asserted. That job hasn’t changed. It’s just gotten harder, and it matters more than ever that we get it right.