KPMG Australia partner fined A$10,000 for cheating on internal AI training exam

KPMG Australia has fined an unnamed partner A$10,000 after internal monitoring found the partner had uploaded a course reference manual into an external AI platform to generate answers on an internal training assessment about AI. The partner, described in reporting as a registered company auditor and an audit partner, was required to retake the test after an internal investigation led to the disciplinary action.

The firm’s monitoring detected the activity after the partner completed the AI training in July and the uploading was flagged in August, according to the timeline provided by KPMG’s internal systems. KPMG’s investigation concluded the partner breached firm rules by inputting the recommended reference document into an external AI tool to answer an exam question, and the firm imposed the financial penalty and the re-test requirement as corrective measures.

IAB reporting indicates the fine was structured as a deduction from future income, describing the sanction as “more than A$10,000” in some accounts; other coverage has used the headline figure A$10,000 and provided varying currency conversions. The partner subsequently self-reported the matter to Chartered Accountants ANZ, which is now examining the case, and remains unnamed in the firm’s disclosures.

KPMG has identified a broader pattern of internal exam misuse since July, registering 28 cases where staff used AI to assist with internal assessments. The firm has been strengthening its AI-detection capabilities after earlier waves of test misconduct, and has upgraded monitoring and controls in the wake of widespread cheating on internal tests between 2016 and 2020 and a 2021 penalty that led to a A$615,000 fine related to improper answer-sharing involving more than 1,100 partners.

Andrew Yates, chief executive of KPMG Australia, acknowledged the scale of the challenge inside the firm, saying, “Like most organisations, we have been grappling with the role and use of AI as it relates to internal training and testing. It’s a very hard thing to get on top of, given how quickly society has embraced it.” Yates added, “Given the everyday use of these tools, some people breach our policy. We take it seriously when they do. We are also looking at ways to strengthen our approach in the current self-reporting regime.”

KPMG has said it will separately report AI-related cheating cases in its annual results and intends to verify that staff have complied with their duty to self-report misconduct to professional bodies. The firm’s push to track AI misuse comes as it expands AI-related advisory work, a growing component of its global business that has been cited in coverage as part of a A$57 billion revenue footprint.

The case sits alongside regulatory shocks across the profession in recent years, including multijurisdictional fines for exam misconduct and a 2022 sanction on KPMG UK. City-level reporting has also recounted failures by other firms, such as a partial refund by Deloitte Australia after an assurance report contained AI-related errors, a matter a government spokesperson acknowledged by saying, “Deloitte conducted the independent assurance review and has confirmed some footnotes and references were incorrect.” KPMG’s internal follow-up, the Chartered Accountants ANZ review, and the firm’s planned annual disclosure are the next developments to watch.