PwC caught publishing AI hallucinations in Middle East reports with fake footnotes

PwC published 'thought leadership' reports marred by AI hallucinations

PwC published reports on AI and electric vehicles riddled with fake footnotes, misattributed claims and unverifiable information, the latest example of a Big Four firm's slapdash use of AI-generated content. The AI hallucinations were contained in "thought leadership" reports designed to drum up consulting work for partners in the Middle East, according to an investigation by researchers at GPTZero verified by the FT. The discovery of AI-generated errors risks embarrassing consulting firms such as PwC that are marketing their services as advisers to companies adopting the technology, including on how to use it responsibly and implement policies to avoid errors. GPTZero's earlier investigations led Big Four rivals EY and KPMG to retract reports that also appeared to include hallucinations. The research group identified four reports PwC Middle East produced over the past two years whose text appeared to have been written with particularly heavy help from AI and whose footnotes revealed shaky foundations. The reports included a playbook for corporate use of "agentic" autonomous AI bots, a guide for governments on how to improve public services and market predictions for electric and autonomous vehicles across the region. Many of the citations in the reports are problematic in some manner, GPTZero researchers found. Often they link to web pages that do not contain the evidence they purport to and some to pages that do not exist. An academic paper on air quality in Riyadh appears to have been hallucinated entirely by AI, because there is no trace of the study in the journal referenced or by the authors to whom it is attributed. Several more footnotes refer to unusual sources. For example, a teenage blogger with 280 followers on Medium is cited as PwC's source for information about a JPMorgan initiative it calls a "real world success story" of agentic AI. The initiative -- in which the bank automated the review of commercial loan agreements, saving hundreds of thousands of hours of human work -- was first reported in 2017, five years before ChatGPT launched the era of generative AI. PwC Middle East told the FT it "takes the accuracy of our published research seriously and is updating a limited number of supporting citations" in the identified reports. "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, without addressing how the errors were included in the reports. In the past few years, the Big Four have pumped out hundreds of thought-leadership pieces on AI to help attract clients, while also exhorting staff to use the technology themselves to speed up and improve their work. The chaotic signposting of source material in the four PwC Middle East reports is symptomatic of AI-generated research, said Paul Esau, researcher at GPTZero. In one tell-tale sign, a claim that human error is responsible for 90 per cent of traffic accidents is mentioned in three different places in one report, Esau said, the first with a footnote, the second without, and the third with two footnotes -- each footnote citing a different source. "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. One report fails to properly cite other PwC work. It references a PwC survey of Middle Eastern chief executives, of whom 70 per cent said generative AI will significantly affect their business, but the accompanying footnote links to a media report that makes no mention of the survey. Another footnote betrays the use of AI. The URL for a media report on cyber security threats to the energy sector includes "utm_source=chatgpt.com".

PwC is the latest Big Four firm caught publishing AI slop

PwC published four Middle East reports stuffed with fake footnotes and made-up sources, the telltale signs of AI hallucination, and one footnote URL literally ended in "chatgpt.com." It is the fourth Big Four firm caught after KPMG, EY and Deloitte, and it happened the same week PwC's own US chief was warning companies not to misuse AI. The world's biggest consultancies are selling companies expensive advice on how to adopt AI without getting burned. Their own reports keep getting burned. PwC is the latest to be caught. Four of its Middle East "thought leadership" reports were a mess, the Financial Times reported. The reports, on AI and electric vehicles, carried fake footnotes, misattributed claims and made-up sources. The research group GPTZero flagged the errors as AI hallucinations, and the FT verified them. A footnote that ended in 'chatgpt.com' The giveaways were not subtle. One report cited a PwC survey claiming 70% of Middle East chief executives expect generative AI to reshape their business. The footnote linked to a news article that never mentioned the survey. Another footnote's web address still carried the tag "utm_source=chatgpt.com." "The chaotic signposting of source material is symptomatic of AI-generated research," GPTZero's Paul Esau told the FT. This is a firm that markets itself as an adviser on using AI responsibly, including how to avoid exactly these errors. All four, now PwC is not alone. It completes an unwanted set. GPTZero's earlier work pushed KPMG to pull an October report. It had made false claims about AI use at UBS, the UK's NHS and Transport for London. EY retracted a study last month over fake footnotes. Deloitte has twice slipped AI fabrications into government reports, including a $1.6m health plan in Canada that cited papers which do not exist. The pattern repeats each time. The Big Four have pumped out hundreds of AI thought-leadership pieces to win clients, while pushing their own staff to use AI to work faster. The result is AI slop with a blue-chip logo on it. Do as we say The timing is almost too neat. The same week, PwC's US chief executive Paul Griggs was telling Business Insider where companies go wrong with AI. One trap he named was bolting it onto broken processes. "All AI is going to do for you is tell you how bad your messy process really is," Griggs said. His own firm just proved him right. Marking their own homework The awkward part is what it says about the product. These firms charge premium rates for judgment and rigour. Yet no one appears to have checked whether the sources were real. It is the same gap that lets AI-written code ship with security holes: the output looks right, so nobody looks closer. It also cuts against the sales pitch. Consultants are among the loudest voices telling clients AI will transform their work. Their own reports show how it can quietly corrupt it, and how the hype outruns the safe use. The hallucinations are the tell, and a detector keeps finding them. The firms keep publishing anyway.

PwC Joins the Race of AI Mistakes

An investigation found that PwC Middle East's AI reports contained fabricated references, raising fresh concerns about how businesses use AI and why human fact-checking still matters. PwC Middle East is under scrutiny after an investigation found that several AI-assisted reports contained references that did not exist. The issue came to light after readers tried to verify the sources and found that many could not be traced. Ironically, these reports were made to advise businesses and governments on AI, autonomous agents and the future of electric vehicles. According to research by GPTZero, reports were , fake citations, broken links, and academic studies that never existed. During the research, the researchers found, in one instance, the source of the URL says, 'utm_source=chatgpt.com.' PwC Middle East said it is reviewing the matter and is fact-checking each of its reports. Previously, GTPZero has caught similar AI-generated reports published by KPMG and EY, which had . The case has once again highlighted one of AI's biggest weaknesses. These systems can produce fluent and convincing text, but they can also invent facts or references if their answers are not carefully checked. This is why many experts say AI should support human work, not replace it.