WHY CONSULTANCIES LOVE AND HATE AI

Clients want to know how much of the work they pay a fortune for has been done by bots

- Devina Sengupta

The Australian winter, which runs from June to August, proved to be rather harsh for Deloitte, one of the Big Four accountancy and consultancy firms. Not because of climate change, but because of artificial intelligence (AI).

Deloitte's member firm in Australia had been tasked with reviewing the country’s welfare system, which had been facing flak for wrongly penalizing welfare recipients and jobseekers. The Deloitte team submitted a 237-page review last December. After the department of employment and workplace relations published the review on its website in July, University of Sydney academic Christopher Rudge pointed out that it was riddled with errors. And then, hell broke loose for Deloitte.

It turned out that Deloitte had relied on AI to compose the report. The AI hallucinations that followed, including mentions of nonexistent academic research papers as well as a fabricated quote from a court judgment, sparked an uproar in Australia.

They “misused AI and used it very inappropriately: misquoted a judge, used references that are nonexistent... I mean, the kinds of things that a first-year university student would be in deep trouble for,” thundered Australian senator Barbara Pocock during an interview with the Australian Broadcasting Corporation, coming down hard on the incompetency displayed by the consultancy in the report.

Subsequently, a revised version was quietly published. After reviewing its review, Deloitte had admitted that “some footnotes and references were incorrect,” the department of employment said in a statement on 3 October. The new version disclosed that Azure Open AI, a generative AI (GenAI) language system, was used to create the report. The report published in July had no such disclaimer.