WA Government Seeks Deloitte Refund Over Flawed University Merger 2026

The Western Australian government is seeking an expedited cost recovery from Deloitte after discovering significant errors in a university merger report the firm produced. The report, which cost taxpayers $166,000, contained AI-hallucinated references and a $260 million cost discrepancy, prompting the government to express it is 'extremely disappointed.'

The feasibility study examined potential mergers between the University of Western Australia (UWA), Curtin University, and Murdoch University. It outlined three merger options with low, mid, and high cost scenarios. However, a close examination revealed that the numbers simply didn't add up, undermining the report's credibility.

What Went Wrong in the Deloitte University Merger Report

The most glaring error appears in the high cost scenario for a merger between Curtin and UWA. The report calculated 'total costs to day 1' at $514 million and 'post day 1' costs over five years at $145 million. The sum of these two figures should be $659 million, but the report stated $399 million—a discrepancy of $260 million.

In addition to the arithmetic mistake, the report contained references that appear to have been generated by artificial intelligence, so-called 'hallucinations' where AI produces false information. These fabricated references further erode trust in the document's reliability.

Professor Stephen Bartos, an expert in public sector governance at the University of Canberra, said the errors are so extensive that a new report may be necessary. 'In my view, they're so extensive that you probably need a new report, or at the very least someone to go through that material from the ground up and rework it manually without the use of AI,' he said. He added that any errors in a report on such a controversial topic discredit the report and that Deloitte should not have allowed it to be published in that state.

Why the Errors Matter for Public Trust

The WA government commissioned the report to inform a high-stakes decision about the future of three major universities. When a report contains basic arithmetic errors and fabricated citations, it undermines the evidence base for any policy decision. This is particularly sensitive because university mergers are controversial and can affect thousands of students, staff, and communities.

The government's demand for reimbursement signals a growing intolerance for poor-quality consultancy work. It also raises questions about quality assurance processes at major firms like Deloitte, especially when AI tools are used in research and analysis.

Professor Bartos's comment that the material should be reworked 'manually without the use of AI' highlights a broader concern: the inappropriate use of AI in critical public-interest work. While AI can boost efficiency, it must be carefully supervised, especially in reports that inform government policy.

The Cost Discrepancy: A Breakdown

To understand the scale of the error, consider the high cost scenario for a Curtin-UWA merger. The report's appendix shows two cost components: $514 million to day 1 and $145 million post day 1 over five years. The correct total is $659 million, yet the report states $399 million. This $260 million gap could have led policymakers to underestimate the financial impact of a merger significantly.

Such a discrepancy is not a minor rounding error; it is a fundamental mistake that could alter the conclusion of a feasibility study. It also raises doubts about other figures in the report that may not have been independently verified.

What This Means for Deloitte and the Consultancy Industry

Deloitte is one of the world's largest professional services firms, and its reputation rests on accuracy and reliability. This incident could damage its standing with government clients, at least in Australia. The WA government's pursuit of cost recovery is a clear signal that clients will hold consultants accountable for substandard work.

Moreover, the use of AI in consultancy deliverables is under increased scrutiny. Firms must ensure that AI-generated content is thoroughly checked and that references are verified. The hallucinated references in this report suggest that proper verification did not occur.

The incident may also prompt other governments to review their procurement and quality assurance processes for consultancy reports. If a firm of Deloitte's stature can produce such flawed work, smaller firms may face even greater scrutiny.

Next Steps: What to Watch

The WA government has requested an expedited cost recovery, meaning it wants its money back quickly. Deloitte has not yet publicly responded to the demand. It remains to be seen whether the firm will refund the $166,000 or contest the request. The background to this story is set out in UK Super-University Merger: Greenwich & Kent Launch in September. A related thread runs through UK's First Super-University: Greenwich and Kent Merger Launches. Earlier coverage in UK's First Super-University Merger: Greenwich and Kent Launch shows how the argument developed.

Additionally, the government may need to commission a new report or have the existing one reworked, as Professor Bartos suggests. That would add further costs and delay decisions about the university mergers. The controversy could also affect public opinion on the proposed mergers, which are already contentious. A closer look at the same subject appears in UK's First Super-University Merger Launches in September. Anyone following the wider picture can start with UK's First Super-University: Greenwich and Kent Merge. The same theme surfaces in Australia: latest articles, which takes a different angle.

Finally, this case underscores the importance of human oversight in AI-assisted work. As AI tools become more common in research and analysis, the need for rigorous quality control becomes even more critical. The WA government's experience serves as a cautionary tale for organizations relying on AI-generated content without proper checks. Context on how this played out before sits in Jason Arday Affair Exposes Academia's Flawed Appointment Systems.