Deloitte has agreed to partially refund the Australian government after acknowledging that an official report it produced contained multiple factual and citation errors caused by artificial intelligence, according to a statement from Australia’s Department of Employment and Workplace Relations.
As reported by The Financial Review, the Big Four consultancy admitted that sections of the A$439,000 (US$290,000) “independent assurance review” were created using a generative AI system, which led to incorrect footnotes and references — including citations to non-existent academic studies from the University of Sydney and Lund University in Sweden.
The department commissioned Deloitte in December last year to assess issues with Australia’s welfare system, which had been criticized for automatically penalizing jobseekers. The review, first released earlier in 2024, was quietly replaced with a corrected version uploaded on Friday after the errors came to light in late August.
Deloitte has now agreed to repay the final instalment of its contract. “The updates made in no way impact or affect the substantive content, findings and recommendations in the report,” the firm said in the amended version. Still, the Australian government described the incident as “embarrassing” and said the corrected document would be published publicly once the refund process is finalized.
The fiasco has drawn attention to the growing risks of using AI tools in professional consulting, particularly the problem of “hallucinations” — when AI systems generate plausible but false information. In its revised report, Deloitte confirmed that it had used a “generative artificial intelligence (AI) large language model (Azure OpenAI GPT-4o)” to assist with parts of the document, though it did not explicitly attribute the mistakes to the model.
This episode adds to mounting concerns over AI reliability in the corporate sector, especially among major firms that are aggressively adopting automation. The UK’s accounting regulator recently warned that the Big Four — Deloitte, PwC, EY, and KPMG — are failing to monitor how automated tools and AI affect audit quality, even as they scale up their use of such technologies for risk assessments and data analysis.
Deloitte, like its rivals, has invested billions in AI development to speed up research and analysis for clients while cutting costs. However, the Australian case illustrates how overreliance on AI without proper verification can damage credibility and client trust.
In a brief statement, Deloitte Australia said, “The matter has been resolved directly with the client.”
While the substance of the report remains unchanged, the episode stands as a cautionary tale for the global consulting industry — and a reminder that even the most sophisticated AI systems can make simple but costly mistakes.
