PwC reports may have AI-generated content; study finds fake footnotes and claims

PwC, one of the world’s Big Four accounting and consulting firms, has been accused of publishing reports with fake references, wrong citations and information that could not be verified. A new study says many of these mistakes appear to have been caused by the use of artificial intelligence (AI).

The study was carried out by AI detection company GPTZero, which reviewed four PwC Middle East reports published between 2024 and 2026. The reports were about topics such as artificial intelligence, electric vehicles (EVs), self-driving cars and improving government services, according to the Financial Times.

According to GPTZero, the reports contained fake footnotes, references that did not match the claims, links to web pages that did not exist and information that researchers could not verify. In one case, a research paper mentioned in a report could not be found anywhere, which some believed it may have been completely made up by AI.

The researchers also found that one report used a blog written by a teenager on Medium as a source to explain a JPMorgan AI project. Researchers questioned this because such an important example should have been supported by a trusted source. They also pointed out that JPMorgan’s AI project was from 2017, before ChatGPT and other generative AI tools became popular.

GPTZero researcher Paul Esau said the reports showed signs of AI-generated writing. For example, one report repeated the claim that 90% of road accidents are caused by human error three times within two pages, but each time it used different references or no reference at all.

“This claim isn’t fake, but no human is going to cite the same fact three times in two pages using three different sources,” he said.

Another report claimed that 70% of Middle East CEOs believe generative AI will significantly affect their businesses, but the footnote linked to a news article that did not mention the survey. Researchers also noticed that one reference link contained “utm_source=chatgpt.com” in its web address.

Responding to these reports, PwC Middle East said that it takes mistakes in its reports seriously and is correcting some of the references and citations that were found to be wrong.”It takes the accuracy of our published research seriously and is updating a limited number of supporting citations in the identified reports,” the company told the FT.

“Consistent with our approach to responsible AI, we have quality control processes for research and content development we expect all our people to adhere to,” it added.