Only 40% of AI-cited statistics trace back to a study (trakkr.ai)

A recent investigation examined the reliability of statistics cited by five prominent AI assistants—ChatGPT, Perplexity, Gemini, Claude, and Google’s AI Overviews—revealing that only 40% of AI-cited statistics could be traced back to an actual study. The exercise involved asking 200 common questions that required numerical answers and following the citation trails provided by each AI. Alarmingly, many citations led to secondary sources like company blogs or statistics roundups rather than original research, with over half of the references lacking direct connection to verified studies. This finding is significant for the AI/ML community as it underscores the critical importance of source verification in AI-generated responses. The substantial reliance on unverified statistics can mislead users and propagate inaccuracies, often with multiple assistants providing the same unsupported figures, giving a false sense of credibility. As AI tools become increasingly integrated into decision-making processes, the need for clear sourcing methodologies and a commitment to transparency in data representation becomes paramount, pushing developers to improve how these models handle and present statistical information.