AI Remixes, Colorizations & Copyright: Who Owns a Machine-Altered Masterpiece?

AI Fabricated Quotes in a Book About Truth — The Irony, the Fallout, and What It Means for Creators

The book was called The Future of Truth. The quotes inside it were fake. And the author says he'll keep using AI anyway.

In May 2026, journalist and author Steven Rosenbaum found himself at the center of a publishing scandal that reads like a cautionary tale written by an AI itself. His new book, The Future of Truth: How AI Reshapes Reality, promised to explore how artificial intelligence was "bending, blurring, and synthesizing" truth. Instead, it became the story: at least six citations in the book, including three entirely fabricated quotes, were traced back to AI hallucinations.

The irony is almost too perfect. But behind the headlines lies a serious question that every writer, publisher, and content creator now faces: when AI tools fabricate content that ends up in published work, who's responsible — and what are the legal consequences?

What Happened: The Future of Truth Scandal

The controversy broke on May 19, 2026, when The New York Times published an investigation revealing that Rosenbaum's book contained what he later acknowledged as "a handful of improperly attributed or synthetic quotes."

Among the findings:

• A quote attributed to tech journalist Kara Swisher that she told The Times she "never said"
• Passages attributed to Northeastern University professor Lisa Feldman Barrett that "don't appear in [my] book, and they are also wrong"
• Three "synthetic quotes" with no identifiable source whatsoever — pure AI fabrication
• Six problematic citations out of 285 total — a small percentage, but the kind of error rate that destroys credibility

Rosenbaum, to his credit, took responsibility. He announced a full "citation audit" with his editors at Simon & Schuster and pledged to correct future editions. But his response to the broader question — will you stop using AI? — was what truly set the publishing world abuzz.

"The idea of taking X years off [from AI] while it sorts itself out, and going back to, like, Microsoft Word … it's just not in my nature," he told Ars Technica. He called AI "magical" and "intoxicating," while also admitting: "It's strangely creative and crafty and unusual in all these ways … and then it betrays you in ways that are just really quite horrible."

How It Happened: The AI-Assisted Workflow That Failed

Rosenbaum's process provides a window into how AI tools are being used — and misused — in professional publishing today.

According to his own account, the workflow looked like this:

1. AI Research: Rosenbaum used OpenAI's ChatGPT and Anthropic's Claude to "surface ideas, locate articles, summarize themes, identify people or papers I might want to look into"

2. AI-Generated Nuggets: Any fact or quote pulled by AI was tagged with a "this came from AI" warning in his notes

3. Human Writing: Rosenbaum says the "actual reporting, narrative structure, interviews, arguments, and conclusions" were entirely his own

4. Professional Fact-Checking: A fact-checker and two copy editors provided by Simon & Schuster reviewed the manuscript

And yet, fabricated quotes still made it into the final book. Why?

The answer points to a systemic problem: traditional fact-checking processes are not designed to catch AI hallucinations. Human fact-checkers verify that a quote exists in a source. But when an AI generates a quote that sounds plausible — attributed to a real person who writes about that topic — a fact-checker may not think to verify that the person never actually said it. You can't check for something you don't suspect is fake.

The Copyright and Legal Implications

The Future of Truth scandal raises several significant legal questions at the intersection of AI and copyright law.

1. Fabricated Quotes and Defamation Risk

When an AI generates a quote and attributes it to a real person, and that quote ends up in a published book, the author and publisher could face defamation claims. Even if the fabricated quote isn't explicitly damaging, attributing false statements to a real person can harm their reputation — and under U.S. defamation law, "reckless disregard for the truth" can meet the standard for liability.

Kara Swisher, Lisa Feldman Barrett, and others whose names appeared alongside fabricated quotes could theoretically pursue legal action. The fact that the quotes came from an AI rather than being deliberately invented by Rosenbaum may not be a complete defense — especially given Rosenbaum's acknowledged awareness of AI's tendency to hallucinate.

2. Copyright in AI-Assisted Works

This case adds fuel to an already heated debate: how much AI involvement is too much for copyright protection?

The U.S. Copyright Office has been clear: works created entirely by AI cannot be copyrighted. But works that involve AI assistance — where a human exercises sufficient creative control — may be eligible. The question is where the line is drawn.

Rosenbaum insists AI didn't write the book. But if AI generated research materials that shaped the book's content, at what point does the human authorship become diluted? The Copyright Office's 2025 AI Report Part 3 addressed training data issues but left many questions about AI-assisted authorship unresolved.

3. Publisher Liability and Due Diligence

Simon & Schuster now faces uncomfortable questions about its editorial standards. The publisher provided a fact-checker and two copy editors — standard practice for non-fiction — yet fabricated content still slipped through. This raises the question of whether publishers need to develop new AI-specific verification protocols beyond traditional fact-checking.

Some potential measures publishers may need to adopt:

• Mandatory AI disclosure: Authors must disclose which tools they used and how
• Source verification tracking: Every AI-sourced claim must be traced back to a verifiable primary source
• AI hallucination audits: Systematic checking of all quotes attributed to living persons
• Contractual warranties: Authors warrant that AI-generated content has been independently verified

4. The FTC and Consumer Protection

If a book marketed as non-fiction contains fabricated quotes — even if accidentally generated by AI — could that trigger consumer protection scrutiny? The FTC has increasingly taken an interest in AI-related deception. While it hasn't yet targeted book publishers, the Rosenbaum case illustrates how easily AI can introduce falsehoods into products consumers expect to be truthful.

The Broader Trend: AI Hallucinations in Publishing

The Future of Truth scandal is not an isolated incident. It's part of a growing pattern:

• Legal filings with fake citations: In 2023, two New York lawyers were sanctioned after ChatGPT generated fake case citations that they submitted to a federal court
• Academic papers with AI-hallucinated references: Multiple journals have retracted papers after discovering AI-generated citations that don't exist
• News articles with fabricated sources: Several outlets have been embarrassed by AI-written articles containing invented quotes and statistics
• AI-generated content in self-publishing: Amazon's Kindle store has seen a flood of AI-generated books, some containing factual errors and fabricated content

A February 2026 study published in Nature found that AI detection tools remain unreliable, making it difficult for publishers and platforms to automatically flag AI-generated or AI-influenced content.

What Creators Should Learn from This

If you're a writer, journalist, content creator, or publisher using AI tools, the Rosenbaum case offers several hard-won lessons:

1. AI Is a Research Assistant, Not a Researcher

Treat AI outputs the way you'd treat tips from an enthusiastic but unreliable intern. They might point you in the right direction, but you must verify everything independently. Never cite an AI-generated quote, statistic, or fact without tracing it back to a verifiable primary source.

2. Fact-Checking Needs an AI Upgrade

Traditional fact-checking assumes that if a quote exists somewhere, it can be found. AI-generated quotes don't exist anywhere — they're invented whole cloth. Fact-checkers need to add a new step: for every quote attributed to a living person, contact that person or their representative for verification.

3. Disclosure Is Not Optional

If you use AI in your creative process, disclose it. The Copyright Office's guidance, the EU AI Act's transparency requirements, and basic professional ethics all point in the same direction: readers have a right to know when AI has been involved in creating the content they consume.

4. Contracts Need AI Clauses

Publishers and platforms should update their contracts to address AI usage explicitly:

• Require disclosure of AI tools used
• Warrant that all AI-generated content has been independently verified
• Include indemnification clauses for AI-related errors
• Define acceptable and unacceptable AI use

5. The "Human in the Loop" Is Only as Good as the Process

Rosenbaum had fact-checkers and editors. The fabricated quotes still got through. Simply having humans review AI output isn't enough — the review process itself needs to be redesigned for an AI era.

The Bigger Question: Can We Trust AI-Assisted Non-Fiction?

At its core, the Future of Truth controversy forces a confrontation with an uncomfortable question: if AI can insert plausible-sounding falsehoods into carefully fact-checked non-fiction, can readers trust any AI-assisted work?

The answer is nuanced. AI tools can genuinely enhance research — surfacing relevant sources, identifying patterns across large datasets, and helping writers explore angles they might otherwise miss. But they can also generate convincing falsehoods that are hard to catch.

The solution isn't to abandon AI, but to build better verification systems around it. Some promising approaches:

• Source-linked AI tools that only cite from verified databases
• Blockchain-based content provenance tracking
• AI hallucination detection tools specifically designed for publishing workflows
• Industry standards for AI-assisted non-fiction, similar to the CAA (Content Authenticity Initiative) standards for images

What's Next for Rosenbaum and The Future of Truth

Rosenbaum is working with Simon & Schuster on corrections. The "citation audit" will likely result in errata and revised editions. Whether the controversy boosts or damages book sales remains to be seen — but one thing is certain: The Future of Truth has become a case study in the very phenomenon it set out to document.

For Rosenbaum personally, the experience hasn't dampened his enthusiasm for AI. "There was never a time when AI was writing the book," he maintains. But his admission that he finds AI simultaneously "intoxicating and dangerous" captures the paradox facing every creator in 2026.

Key Takeaways

• AI tools fabricated at least six citations — including three entirely invented quotes — in Steven Rosenbaum's book The Future of Truth
• Traditional fact-checking processes failed to catch the fabrications because they weren't designed to detect AI hallucinations
• The case raises significant questions about defamation liability, copyright eligibility for AI-assisted works, and publisher due diligence
• Creators using AI should treat all AI-generated claims as unverified tips, redesign fact-checking workflows, and always disclose AI involvement
• Readers, publishers, and platforms need new standards for verifying and trusting AI-assisted non-fiction
• The incident underscores why AI transparency and content provenance are becoming critical issues in copyright law

---

This article is for informational purposes only and does not constitute legal advice. For specific legal questions about AI use in publishing, consult a qualified attorney.

Published by AI Copyright Legal. We cover the evolving intersection of artificial intelligence and copyright law.