Nonfiction Book 'The Future of Truth' Reportedly Included AI-Generated and Misattributed Quotations

Steven Rosenbaum's nonfiction book The Future of Truth reportedly included AI-generated or misattributed quotations after he used ChatGPT and Claude while working on the book. The New York Times reported that more than half a dozen quotations in reviewed sections were synthetic or wrongly attributed, including passages assigned to named writers and scholars. Rosenbaum later acknowledged the errors as accidental and said affected passages would be reviewed for correction.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.1 False or misleading information

The AI systems generated false and fabricated quotes that were published as factual, leading to the spread of misleading information in a nonfiction book.

Additional risk subdomains

• 5.1 Overreliance and unsafe use: The author over-relied on ChatGPT and Claude for research and verification without independently fact-checking the generated quotes.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The fabrication of quotes was an unintentional outcome of the AI systems' generation process during post-deployment use by the author.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The incident involves chatbots and AI-generated content, which pose a moderate risk and require transparency and verification to prevent the spread of misinformation.

AI system and alleged parties

• AI system: ChatGPT, Claude (Anthropic, OpenAI)
• AI purpose: Writing Assistant; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI, Anthropic
• Alleged deployer: Writers, Steven Rosenbaum, Authors
• Alleged harmed parties: Readers, People misquoted by AI-assisted books, Meredith Broussard, Lisa Feldman Barrett, Lee McIntyre, Kara Swisher, Epistemic integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Epistemic

Reported: The report explicitly describes epistemic harm, specifically the fabrication of quotes and misattributions in a published nonfiction book.

Directly caused: The AI tools generated several fake or misattributed quotes that were published in the book 'The Future of Truth'.

Indirectly caused: The publication of these fake quotes misled readers and undermined the credibility of the book, the author, and the quoted experts.

Inferred additional harm: Readers of the book and the Wired excerpt were exposed to fabricated information, potentially corrupting their understanding of the subjects discussed.

People affected

• Occurrences reported: 1
• People reportedly harmed: 5
• People reportedly exposed: 1005

Potential causes

Management

• Failure in Publisher Oversight: Publishers did not implement strict compliance checks for AI usage.
• Lack of AI Disclosure Policies: No formal management policies requiring verification of AI-generated inputs.

Technology

• AI Hallucination of Quotes: ChatGPT and Claude generated realistic but entirely fabricated quotes.
• Inaccurate Source Attribution: AI models mixed real quotes with incorrect source origins.

Data Inputs

• Lack of Real-Time Data Access: AI models trained on static data without live verification links.
• Pattern-Based Text Generation: Models generated plausible-sounding text instead of factual quotes.

Human Factors

• Overreliance on AI Output: Author trusted AI-generated research without independent verification.
• Lack of Fact-Checking Diligence: Author failed to manually verify quotes before publication.

Process and Methods

• Inadequate Editorial Review: Publisher lacked rigorous verification processes for AI-assisted text.
• Unverified AI Research Workflow: Integration of AI in writing lacked structured validation steps.

Regulatory Environment

• Absence of AI Publishing Standards: No industry-wide regulations or guidelines for AI-assisted books.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, firsthand accounts from the author admitting to the use of AI and the resulting errors, as well as direct confirmations from the misquoted individuals.

An author's unintentional inclusion of AI-generated fabricated quotes in a nonfiction book highlights the risks of overreliance on AI tools for research. The incident represents a minor epistemic issue within the publishing industry with negligible implications for national security.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: No clear primary
• Alleged perpetrator: PERSON_001

Threat characteristics

• Imminence: Long-term. This is a strategic concern regarding AI-generated misinformation in media and publishing rather than an immediate national security crisis.
• Autonomy: Human-controlled. The AI acted as a writing assistant under human control, though the author failed to adequately verify the outputs before publication.
• Novelty: Established threat. AI hallucinations and the generation of fabricated quotes are well-documented and established limitations of large language models.

Impact by dimension

• Physical security: Negligible. No physical security systems, critical infrastructure, or human safety assets were affected or targeted in this incident.
• Information security: Negligible. The incident involves unintentional publishing errors rather than coordinated state-sponsored disinformation, model theft, or intelligence compromise.
• Sovereignty: Negligible. No government functions, electoral systems, or state sovereignty operations were impacted by the book's publication.
• Economic security: Negligible. There was no threat to strategic industries, financial systems, or national technological advantages.
• Societal stability: Negligible. While the incident caused minor epistemic harm to readers, it does not represent a large-scale threat to social cohesion, civil liberties, or human rights.