Global consulting firm Deloitte is refunding a portion of its $440,000 fee to the Australian government after using generative AI in a report on the “Future Made in Australia” initiative. Commissioned in 2024, the report reviewed the government’s compliance framework and IT system that penalizes job seekers.
The report, released in July, contained significant errors, including non-existent academic citations and a fabricated Federal Court quote, according to the Australian Financial Review. The Department of Employment and Workplace Relations has since published a revised version, correcting these errors and removing fictitious references.
Dr. Christopher Rudge, an Australian welfare academic who identified the issues, attributed the errors to AI “hallucinations.” The revised report replaced the AI-generated citations with multiple new, valid sources, suggesting the original claims lacked evidentiary basis.
Deloitte stated that AI was used only in the initial drafting stages and that human experts extensively reviewed the final product. The firm maintains that AI did not affect the report’s content, findings, or recommendations. The company also disclosed the use of Azure OpenAI GPT-4o in its research.
A Deloitte spokesperson confirmed resolution with the client, and the Department confirmed the refund and potential stricter AI usage guidelines for future contracts.
This incident has ignited discussions about the ethical and financial implications of AI in consultancy work. As firms like Deloitte, which recently partnered with Anthropic to provide its employees access to the Claude chatbot, increasingly adopt AI, questions arise about human oversight and value for clients. This case is a major instance where consequences followed undisclosed AI use in Australian government work.
Also Read: OpenAI introduces apps inside ChatGPT
