In a surprising turn, Deloitte has agreed to repay part of a $440,000 government fee after admitting it used AI to write portions of an official report — which later turned out to contain fake quotes and references.
The report, commissioned by Australia’s Department of Employment and Workplace Relations, was meant to review an IT system managing welfare penalties. To fill “traceability and documentation gaps,” Deloitte used OpenAI’s GPT-4o to draft parts of the document.
Fake Citations and Quotes Exposed
After its release in July 2025, a University of Sydney academic uncovered multiple factual errors, including nonexistent research papers, invented citations, and even a fabricated court quote from a welfare case.
The department later confirmed Deloitte would refund the final payment, though it stated that the report’s main findings and recommendations remain unchanged.
A Wake-Up Call for AI in Professional Work
The incident highlights growing concerns about AI hallucinations — when models confidently generate false information. As companies increasingly use AI for research and report writing, experts warn that stronger verification and human oversight are critical before publication.
AI may accelerate productivity, but as this case shows, it can also damage credibility if not carefully monitored.
