Login
iscover
Share this article on:
Powered by
MOMENTUM
MEDIA
Powered by MOMENTUMMEDIA
The report contained a wealth of mistakes, which are believed to have been caused by unchecked AI hallucinations.
Global big four consulting firm Deloitte will partially refund the Australian government $440,000 after the company admitted to using AI to write a report for the government, which ended up being full of mistakes and made-up points.
The report, which was commissioned by the Department of Employment and Workplace Relations (DEWR), was initially published in July, but was quickly identified as having inaccurate footnotes and made-up references and quotes.
The issues with the paper were first reported by The Australian Financial Review journalist Paul Karp, when welfare academic Chris Rudge from the University of Sydney first identified errors in the report after finding over a dozen fabricated points attributed to his colleagues.
“I was in no doubt, when I read the names of the works, that they were fake because I work in the area and I know my colleagues’ work,” he said.
“Once I discovered one, I just discovered more and more,” he continued, adding that he had no doubt that the errors were the fault of AI.
Rudge also identified a fabricated quote from the Deanna Amato v Commonwealth robo-debt case.
The report reads that “Justice Davis” (an incorrect spelling of Justice Davies) said that “the burden rests on the decision-maker to be satisfied on the evidence that the debt is owed. A person’s statutory entitlements cannot lawfully be reduced based on an assumption unsupported by evidence,” in paragraphs 25 and 26.
However, the consent orders issued by Federal Court Justice Jennifer Davies in the 2019 case don’t actually contain a paragraph 25 or 26. Another quote was attributed to paragraph 30, but actually appeared in paragraph 9.
Another part of the report claimed to reference a paper titled The Rule of Law and Administrative Justice in the Australian Social Security System, which doesn’t actually exist.
Another fake article, Choice, Responsibility and the Regulation of Behaviour: Lessons from the Social Security System, which was apparently written by Carolyn Adams and published by the UNSW Law Journal in 2012, was also referenced.
“I cannot understand how a human could create titles of works that don’t exist, that don’t appear on Google. How would they do that?” he said.
“I think it’s good if it’s AI, because to think of a person doing that is almost worse. It’s very disrespectful to those who have done the research to just not get it right.”
Following the wealth of identified issues, Deloitte uploaded a new version of the report, later uploading it again with a note saying that “a small number of corrections to references and footnotes”. Deloitte claims that the mistakes did not impact the report’s findings.
“Deloitte conducted the independent assurance review and has confirmed some footnotes and references were incorrect,” a spokesperson for the DEWR confirmed.
“The substance of the independent review is retained, and there are no changes to the recommendations.”
The consultancy firm, which offers consultancy services to other organisations regarding the best uses and ways to adopt AI technology through its AI Institute, also confirmed that it had used “a generative artificial intelligence (AI) large language model (Azure OpenAI GPT-4o) based tool chain licensed by DEWR and hosted on DEWR’s Azure tenancy” in the report’s creation.
Deloitte has not confirmed whether or not the mistakes were caused by AI generation. However, a spokesperson said that “the matter has been resolved directly with the client”.
Be the first to hear the latest developments in the cyber industry.
