The report, titled Redefining excellence in the age of agentic AI, claimed that organisations such as UBS, the UK’s National Health Service, Swiss Federal Railways and Transport for London had deployed advanced AI tools in specific ways. Researchers at GPTZero reviewed those case studies and concluded several of them were fabricated, a judgment later backed by independent verification from a major financial publication. UBS, after being told its supposed AI use was wrongly described, said it would ask KPMG to strip out the false claims. KPMG, one of the big four professional services firms, subsequently pulled the report from some of its websites on Thursday, which was Friday in Australian Eastern Standard Time.
GPTZero had already become known for scrutinising AI-generated content in corporate and academic settings and its latest findings show how easily polished reports can smuggle in hallucinated detail. The identified inaccuracies stem from AI tools inventing plausible-sounding case studies that went unchecked before publication. Industry observers say firms racing to showcase AI leadership risk leaning on the same technology for research, blurring the line between evidence and fabrication. KPMG withdrew the report only weeks after EY withdrew a separate study when GPTZero uncovered fake footnotes and other serious errors in that document.
Large advisory firms are still struggling to build reliable guardrails around internal AI use. For clients, it raises specific questions about how much trust to place in high-profile surveys and thought leadership on emerging technologies. The recent failures give regulators and professional bodies ammunition to push for tighter standards on how firms source, verify and present data tied to AI. Marketing pitches about cutting-edge innovation are colliding with governance failures inside firms that sell themselves as trusted advisers.
KPMG Whistleblower Clash Exposes Culture Gap
Inside KPMG Australia, a formal whistleblower disclosure meant to trigger legal protections opened a grinding standoff that has dragged on for two years. An executive who tried to use the firm’s much-advertised “speak up” framework discovered that putting concerns in writing did not lead to swift investigation, but to a slow procedural deadlock.
The executive lodged the disclosure on 30 May 2024, addressing it to an eligible recipient within KPMG Australia under the Corporations Act whistleblower provisions. Rather than naming individuals, the document described patterns of alleged misconduct including misuse of confidential client information and questions over the integrity of audit tender processes. Before providing further detail, the whistleblower sought explicit confirmation of statutory protections available under Australian law. KPMG declined to give those assurances and the impasse continues.
Allegations about the handling of confidential client data go to the core of a big four firm’s promise of integrity and independence. Concerns raised about audit tenders touch directly on how KPMG competes for and potentially wins lucrative assurance mandates. Instead of triggering broad internal alarm, the disclosure activated what critics describe as a defensive corporate posture. Public messaging about encouraging staff to report issues confidentially sits awkwardly with a response that leaves a whistleblower in limbo.
The clash is part of a wider tension facing large professional services firms that promote robust ethics frameworks while managing commercial and legal risk. Staff are told they will be protected if they raise red flags, yet the drawn-out dispute over basic whistleblower protections suggests otherwise. Observers note that cases like this can chill internal reporting, as employees watch how long and fraught the process becomes for those who come forward. KPMG’s handling of the disclosure now sits at the centre of a test of its culture and governance claims.
