How a KPMG Partner Used AI to Cheat on an AI Exam and What It Means for Workplace Ethics

The Irony of the Automated Audit

In a twist that reads more like a satirical tech blog than a corporate press release, a senior partner at KPMG Australia was recently slapped with a A$10,000 fine. The offense? Using artificial intelligence to pass an internal examination that was specifically designed to teach staff how to use artificial intelligence responsibly. The unnamed partner reportedly uploaded confidential training materials into an external AI platform to generate answers for the test, bypassing the very learning process the firm had mandated.

This incident, which surfaced during an Australian Senate inquiry in February 2026, has highlighted a glaring paradox in the modern workforce. As major firms like the "Big Four" push for rapid AI integration to boost efficiency, they are finding that the same tools they sell to clients as "transformative" are being used by their own leadership to cut corners. It is a high-stakes game of cat and mouse where the "cat" is often the very technology the firm is trying to master.

Beyond a Single Bad Actor

While the partner’s fine has grabbed the headlines, the scale of the problem is much larger. KPMG Australia revealed that it has caught more than two dozen employees—28 to be exact—using AI to game internal assessments since July. Most of these cases involved staff at the manager level or below, but the partner’s involvement is particularly damaging given their role as a registered company auditor. Auditors are held to the highest standards of integrity because their work serves as the bedrock of trust in global financial markets.

To catch these digital cheaters, KPMG ironically had to lean on its own proprietary AI detection tools. This creates an "AI arms race" within the office walls: employees use generative tools to save time on mandatory training, while the firm uses detection algorithms to ensure they actually did the work. As reported by The Guardian, this incident has forced the firm to rethink its entire approach to remote examinations, moving toward more robust monitoring and, in some cases, a return to in-person testing.

A History of Integrity Challenges

For KPMG, this isn't the first time "answer-sharing" has caused a PR nightmare. Back in 2021, the firm was fined A$615,000 by the US audit watchdog after a massive investigation found that over 1,100 partners and staff had shared answers on internal tests between 2016 and 2020. At that time, the cheating was manual—colleagues simply swapped emails with the correct answers. In 2026, the problem has simply been automated.

Critics, including Greens Senator Barbara Pocock, have called the recent A$10,000 fine "extremely disappointing," arguing that such a small penalty for a high-earning partner does little to deter future misconduct. "We’ve got a toothless system where con artists get away with so much," she told a parliamentary committee, reflecting a growing public frustration with self-regulation in the professional services industry.

The Rise of Shadow AI in the Office

The KPMG scandal is a symptom of a broader trend known as "Shadow AI"—the unauthorized use of artificial intelligence tools by employees to complete their work. When a company’s training programs are perceived as "billable time drains" rather than valuable education, even senior leaders are tempted to delegate the task to a chatbot.

CEO Andrew Yates admitted that the firm is "grappling" with how quickly society has embraced these tools. The challenge for 2026 is not just making AI work for the business, but ensuring that the human element of "judgment" doesn't disappear in the process. Experts at the Financial Express suggest that firms may need to move away from "tick-a-box" compliance tests and toward more interactive, performance-based assessments that AI cannot easily replicate.

Final Thoughts

The KPMG partner's fine serves as a stark reminder that technology cannot replace character. While AI can draft an email or summarize a report, it cannot possess the professional skepticism required for an audit. As the corporate world enters this "AI-first" era, the true test won't be how well employees can use the tools, but whether they can be trusted to know when not to use them.