The author of a nonfiction book that delves into artificial intelligence’s effects on our understanding of truth has come forward, admitting that his work includes several inaccurate or fabricated quotes generated by AI.
Stephen Rosenbaum, who wrote true future, acknowledged this week that his book contains multiple misattributed quotes created by AI. His admission followed scrutiny from times, which identified the false citations and contacted him for clarification over the weekend.
In a statement released on Monday, Rosenbaum conceded that the book included “several inappropriate and composite quotations.” He also mentioned that he has begun his own investigation into the issue, labeling the inaccuracies as coincidental and asserting he did not aim to fabricate any perspectives during writing.
“As disclosed in the book’s acknowledgments, I utilized AI tools like ChatGPT and Claude throughout the research, writing, and editing phases,” he noted. “This doesn’t excuse the errors, and I take full responsibility. I’m currently collaborating with my editors for a thorough review to rectify the errors. We will amend these in future editions.”
The book was published by the BenBella Books imprint and distributed by Simon and Schuster. While BenBella operates independently from Simon & Schuster, the latter declined to comment when reached for a response.
Rosenbaum serves as the executive director of the Center for Sustainable Media, a nonprofit dedicated to empowering a new generation of media consumers and creators in an increasingly media-driven landscape. He is known for facilitating discussions among media and technology leaders through both in-person events and online interviews.
true future has garnered significant attention since its launch, with excerpts appearing in wired magazine. The book received endorsements from prominent figures, including journalists like Taylor Lorenz, Michael Wolff, and Nicholas Thompson, the chief executive of the atlantic ocean. Maria Ressa, a Nobel Peace Prize-winning journalist critical of former Philippine President Rodrigo Duterte, wrote the foreword.
times identified several fabricated quotes appearing throughout the text. One, attributed to noted tech journalist Kara Swisher, claimed, “The most sophisticated AI language model is like a mirror. It reflects our own morality, elegantly articulated, but ultimately empty beneath the surface. It is not bound by Asimov’s Laws or any ethical framework. It is bound by the patterns of the training data and the goals set by its creator.”
When approached about the quote, Swisher responded via text, stating, “I never said that,” insisting that the quote seemed to be generated by AI rather than originating from him. “I guess I sound better with my head up there,” Swisher added humorously.
In another chapter that discusses social media and fake videos’ effects on teenagers, two quotes are cited. One of them is supposedly from Lisa Feldman Barrett, a psychology professor, who stated, “Emotions are not just reactions to truth; they are also how we construct truth.” However, Barrett clarified to times that this quote is incorrect, explaining that she would never assert that “emotions are not reactions to truth.” She further distanced herself from the notions of integrating emotional and social signals.
Some segments of the book also contain incorrectly attributed real quotes. For instance, a citation from Meredith Broussard’s book, artificial intelligence, was actually taken from her words during a 2023 interview on Marketplace Tech. Broussard remarked that it appeared to be an “AI hallucination or a misquote.”
Rosenbaum suggested in his statement that these errors could highlight important warnings regarding AI in research and validation. “If this episode serves as a caution about the risks of AI-enabled research, then that’s precisely why I wrote the book,” he said. “These AI errors do not detract from the pressing questions this book raises about truth, trust, AI, and their societal impacts.”
