KPMG Australia disclosed that more than two dozen staff used AI to game internal exams since mid-2025, with the cheating detected by the firm’s own AI monitoring tools.

A senior KPMG Australia partner was fined A$10,000 for using generative AI to complete an internal ethics AI training assessment, after which the partner redid the test.

KPMG plans to publicly report AI misuse alongside its annual results to increase transparency and accountability.

Disciplinary proceedings are underway with professional bodies; ASIC will act once those bodies initiate proceedings, noting that audit firms aren’t required to report internal breaches to ASIC.

Critics argued the self-reporting regime is inadequate and urged tighter reporting and greater transparency to address unethical behavior by large consulting firms.

The episode underscores the need for stronger workplace policies and better detection tech to preserve integrity in training, services, and client deliverables.

Observers criticized the firm for prioritizing AI adoption over training design, reflecting broader tensions around AI in professional services.

The incident emerged during an Australian Senate inquiry into industry governance, with Greens senator Barbara Pocock calling oversight ‘toothless.’

Pocock raised the cheating as a misdemeanor during the inquiry, prompting questions to ASIC about enforcement.

The scandal was first reported by the Australian Financial Review and discussed at the Senate inquiry, prompting scrutiny of enforcement.

The case highlights tension between using AI for efficiency and maintaining integrity in training and examinations at major firms.

AI cheating concerns have spilled into sector actions, including ACCA requiring in-person exams and Deloitte Australia facing scrutiny over AI-related issues.