A research report co-authored by Accenture and IE University on artificial intelligence in African agriculture was found to contain false citations, misattributed research and possibly nonexistent case studies. According to Sina Finance, AI detection company GPTZero identified factual errors and 15 citation problems, including inaccurate paper titles and author names.
The report, published in November 2025, cited two seemingly nonexistent organizations under different names and claimed predictive AI detected cacao swollen shoot disease with 99% accuracy in Colombia and Ecuador. Its supporting reference focused on West Africa and did not mention either South American country. Other errors included incorrect rankings for Ghana and Uganda in the University of Oxford’s 2023 AI Readiness Index.
Accenture said its review found the errors resulted from human mistakes and that there was no evidence they were caused by AI hallucinations. The company said it would correct and republish the report, but declined to explain the basis for its conclusion or how the errors occurred. IE University said the authors were reviewing the sources and data and would make necessary corrections if errors were confirmed.
GPTZero co-founder Edward Tian said the errors largely matched patterns typical of AI hallucinations. GPTZero said it manually checked each error it flagged and conducted additional verification using open-source intelligence. The company’s AI detection tools have also faced criticism for sometimes incorrectly labeling human-written content as AI-generated.