Deloitte to Refund Fee Over Australian Government Report Containing Fabricated AI References

Deloitte Australia has agreed to refund a portion of the 440,000 Australian dollars (approximately $290,000) it received from the federal government after a report it produced was found to contain fabricated quotes and references believed to have been generated by artificial intelligence.

The 237-page report, commissioned by the Department of Employment and Workplace Relations, was initially published on the department’s website in July. However, it was swiftly withdrawn and replaced with a revised version after serious errors were discovered.

Concerns were first raised by Chris Rudge, a health and welfare law researcher at the University of Sydney, who found multiple inaccuracies, including a made-up quotation from a federal court judgment and citations to non-existent academic papers. Rudge described the report as “full of fabricated references,” prompting widespread scrutiny of Deloitte’s research practices.

In a statement issued on Tuesday, the Department of Employment and Workplace Relations said Deloitte had acknowledged the inaccuracies and agreed to repay the final installment of its contract. “Deloitte had agreed to repay the final installment under its contract,” the department confirmed, adding that the exact refund amount would be disclosed once processed.

Following the revelations, Deloitte and the department released an updated version of the report on September 26. The revised edition now includes a formal disclosure that a generative AI system, Microsoft’s Azure OpenAI, was used in its creation. It also removes the fabricated material, including the false quote attributed to a federal court judge and several fictitious academic references.

In a section titled Report Update, Deloitte stated that the September version replaces the one published in July and acknowledged that “some footnotes and references were incorrect.” The firm, however, declined to specify whether artificial intelligence had been used in the original drafting of the report.

Responding to questions from the Associated Press, a Deloitte spokesperson said the issue “has been resolved directly with the client,” without further elaboration.

The controversy underscores growing global concerns over the reliability of AI-generated content, particularly in professional and governmental contexts. Experts warn that “hallucinations” — instances where AI systems produce inaccurate or wholly fabricated information — pose significant risks to the credibility of research and policymaking.

The Australian government’s experience with the Deloitte report is among the latest examples of how reliance on AI in professional consulting can lead to reputational and financial consequences, prompting renewed calls for greater transparency in the use of generative technologies.