Incident·WS-0046·Jul 2025·disclosed Oct 7, 2025 (84 days later)·in the wild·Misinformation & fabrication

Deloitte refunds Australia for report with AI-invented sources

Deloitte Australia / Azure OpenAI; Dept. of Employment and Workplace Relations

25/100

WS-2 Moderate

Harm score 20, signal score 31. Autonomy: relied on (people trusted bad output). How the scores work.

Harm 1/5

damage actually done

Reach 1/5

who was affected

Autonomy 1/4

how much the AI did on its own

Novelty 2/5

how new the behavior was

Permanence 1/5

how hard it is to undo

A 237-page Deloitte review for Australia's employment department contained a fabricated quote attributed to a federal court judgment and references to non-existent academic works. After University of Sydney researcher Chris Rudge flagged up to 20 errors, Deloitte reissued it disclosing Azure OpenAI use and agreed to repay the final contract instalment.

Damage

partial refund of an A$440,000 (~US$290,000) contract (amount not initially disclosed)

Sources

Bnnbloomberg. Report first published July 2025; partial refund became public ~6-7 Oct 2025 (disclosed field). Fetched source names Azure OpenAI but not the specific model.

Factor ratings and the score are this index’s judgments by a published formula; facts, dates and figures come from the linked sources. Where a lawsuit is pending, the summary reports allegations, not findings.