This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
An author used AI tools ChatGPT and Claude to write a book about AI's effects on truth, but the AI generated numerous fake or misattributed quotes that were included in the published work without proper verification.
Steven Rosenbaum, author of 'The Future of Truth' published by BenBella Books and distributed by Simon and Schuster, acknowledged that his nonfiction book about artificial intelligence's effects on truth contained more than a half-dozen fake or misattributed quotes generated by AI. Rosenbaum disclosed in his book's acknowledgments that he used ChatGPT and Claude during research, writing and editing. The New York Times discovered the fabricated quotes when they asked sources about statements attributed to them, including technology journalist Kara Swisher, psychology professor Lisa Feldman Barrett, and other experts who confirmed they never made the quoted statements. The quotes appeared throughout various chapters covering topics like AI lies, social media effects, and post-truth concerns. Some quotes were entirely fabricated while others were real quotes incorrectly attributed to wrong sources or publications. The book had received significant attention including an excerpt in Wired magazine and promotional blurbs from prominent journalists. Rosenbaum called the inclusion an accident and took full responsibility, stating he had no intention of fabricating viewpoints and was working with editors to correct future editions.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
