A major KPMG report on AI was found to be chock-full of...AI hallucinations (www.techradar.com)

A recent investigation by GPTZero has uncovered significant issues in a KPMG report on agentic AI, revealing that only five out of 45 citations accurately reflect real sources. The report was found to contain a plethora of AI-generated errors, including completely fabricated citations and misleading case studies, which GPTZero describes as "vibe citing." This term refers to the phenomenon where generative AI produces references that appear credible but are ultimately misleading, raising concerns about the integrity of AI-generated content. This incident is particularly alarming for the AI/ML community because KPMG's reports hold considerable influence and are often widely cited in various discourse, from news articles to academic discussions. The potential for misinformation to spread due to these unreliable citations poses significant risks to researchers and professionals who rely on accurate data. The findings underscore the necessity for vigilance in verifying sources and emphasize that even seemingly plausible AI outputs can perpetuate inaccuracies, prompting a broader conversation about the consequences of AI hallucinations in published works.