Deloitte, a leading professional services and consulting firm, revealed a major AI partnership with Anthropic on the same day news broke that it would be refunding a government client for a report containing flawed AI-generated content.
What this means: Deloitte’s agreement with Anthropic highlights its dedication to artificial intelligence, even as it faces challenges with the technology. This is an issue confronting many organizations, not just Deloitte.
The coincidence of these events is striking—almost ironic. As Deloitte promoted its expanded use of AI, the Australian Department of Employment and Workplace Relations announced that the firm would be issuing a refund for a departmental report that included AI-generated inaccuracies, according to the Financial Times.
The department had paid Deloitte A$439,000 for an “independent assurance review,” which was released earlier this year. In August, the Australian Financial Review revealed that the report contained several mistakes, such as referencing academic papers that do not exist. The revised version was posted on the department’s website last week. The Financial Times reported that Deloitte will return the final payment from its government contract.
TechCrunch contacted Deloitte for a statement and will update this story if a response is received.
On Monday, Deloitte announced it would deploy Anthropic’s chatbot Claude to its global workforce of nearly 500,000 employees. The partnership, which began last year, aims to develop compliance solutions and features for regulated sectors like finance, healthcare, and public services, according to a blog post from Anthropic. CNBC also reported that Deloitte intends to build a variety of AI agent “personas” to serve different company departments, such as accounting and software development.
“Deloitte is making a substantial investment in Anthropic’s AI platform because our philosophies on responsible AI are closely aligned, and together we can transform enterprise operations in the coming decade. Claude remains a top choice for our clients and our own AI initiatives,” wrote Ranjit Bawa, Deloitte’s global leader for technology, ecosystems, and alliances, in the blog post.
The financial specifics of the partnership—which Anthropic described as an alliance—have not been made public.
This arrangement marks Anthropic’s largest enterprise rollout so far and demonstrates how AI is becoming integrated into all facets of daily life, from workplace tools to everyday questions at home.
Deloitte is not the only organization or individual to have recently been caught using inaccurate AI-generated material.
Back in May, the Chicago Sun-Times had to acknowledge that its annual summer reading list included an AI-generated selection of books, after readers noticed that some titles were fabricated, even though the authors were real. According to an internal memo seen by Business Insider, Amazon’s AI productivity tool, Q Business, also faced accuracy issues during its first year.
Anthropic itself has faced criticism for relying on AI-generated information from its own chatbot, Claude. Earlier this year, the company’s attorney apologized after Anthropic used an AI-created citation in a legal dispute with music publishers.
