Steven Rosenbaum Orders Citation Audit After NYT Finds Synthetic Quotes in AI Book
Updated
Updated · Ars Technica · May 22
Steven Rosenbaum Orders Citation Audit After NYT Finds Synthetic Quotes in AI Book
4 articles · Updated · Ars Technica · May 22
A New York Times investigation found a handful of improperly attributed or AI-generated quotes in Steven Rosenbaum’s new book, including lines Kara Swisher said she “never said.”
Rosenbaum said the errors stemmed from his use of AI research tools and is now working with editors on a full citation audit to correct future editions.
The problem undercuts the book’s own warning that fast-moving, profit-driven AI is bending and synthesizing truth, with Lisa Feldman Barrett also disputing a quote attributed to her.
Rosenbaum said he has learned to be far more suspicious of AI outputs, but still does not plan to abandon the tools, calling them “magical” despite the controversy.
If an expert's book on AI's threat to truth contains AI-generated lies, how can readers trust any author who uses AI?
As AI invents 'facts,' who is now liable for falsehoods in books: the author, the publisher, or the AI's creator?
AI Hallucinations in Publishing: The Steven Rosenbaum Controversy and the Urgent Need for Industry Reform
Overview
Steven Rosenbaum's book, 'The Future of Truth,' became the center of controversy after AI-generated fabrications were discovered within its pages. This incident highlighted the risks of relying on artificial intelligence for research, especially the problem of 'AI hallucinations' that can create fake or misattributed quotes. The New York Times led the investigation, revealing specific fabrications and prompting Rosenbaum to acknowledge the issue. The controversy not only questioned the integrity of the book but also sparked a broader discussion about the challenges and credibility concerns facing the publishing industry in the age of AI.