We Asked the ‘Future of Truth’ Author to Explain How He Used AI. It Didn’t Go Well
WIRED pressed Steve Rosenbaum on AI use in his truth-focused book after quote errors surfaced. He gave a broad answer, but not a clean accounting.
Intelligence analysis by GPT-5.4 Mini
After The New York Times reported fabricated or misattributed quotes in Steve Rosenbaum’s The Future of Truth, WIRED reexamined an excerpt and questioned how much AI had been used. Rosenbaum said AI helped with research and editing stages, but he would not clearly say whether passages were copied, pasted, or rewritten from AI output.
A writer made a book about how hard it is to know what is true. Then people found that some quotes in the book were wrong, and that the writer may have used AI tools while making it.
WIRED asked him to explain. He said AI helped him search for facts, think of ideas, and polish words, but he would not clearly say exactly how much he copied or edited from it.
It is a little like a student using a calculator, a notebook, and help from a friend, then trying to explain how much of the homework was really their own work. The big question is where help ends and ownership begins.
Analysis
What happened
WIRED revisited an excerpt from Steve Rosenbaum’s The Future of Truth after the New York Times reported that the book included several made-up or wrongly attributed quotes. WIRED says its own fact-checking had cleared the excerpt before publication, but a reader later flagged it as sounding AI-written.
The AI question
Rosenbaum had acknowledged in the book’s notes that tools like ChatGPT, Claude, NaturalReaders, ProWritingAid, and Grammarly helped refine the presentation of his ideas. WIRED then tested the excerpt with multiple detection tools, which suggested the text was likely AI-generated. Those tools are imperfect, so WIRED went straight to Rosenbaum for a clearer explanation.
Rosenbaum said he used AI in parts of the research and editorial process, including finding sources, brainstorming, getting structure feedback, and polishing language. He insisted the ideas, reporting, and final authorship were his, and denied that the WIRED excerpt was simply AI-generated text published unchanged. But when pressed on whether he had copied and edited AI output into the manuscript, his answers stayed vague.
Why the piece matters
The article is less about one book than about a broader publishing problem: writers are increasingly using AI in ways that are hard to define, hard to disclose, and hard to police. WIRED also uses the episode to stress a second point: detection tools can be suggestive, but they are not proof. The result is a sharp clash between skepticism about AI-assisted writing and the reality that many authors now rely on AI at some stage of the process.
Key points
- WIRED revisited an excerpt from Steve Rosenbaum’s The Future of Truth after reporting on quote errors in the book.
- Rosenbaum said AI tools helped with research, brainstorming, structure, and language refinement.
- He would not give a precise account of whether AI output was copied into the manuscript and edited.
- AI-detection tools flagged the text as likely AI-generated, but WIRED notes those tools can be wrong.
- The piece raises broader questions about disclosure, authorship, and AI use in publishing.
