Siti Muinah
In an era defined by the rapid proliferation of artificial intelligence, the line between authentic insight and algorithmic imitation has become increasingly blurred. Recently, Sam Sifton, the editor behind The New York Times’ widely read newsletter, The Morning, sparked a vital industry conversation with a candid subject line: “Who’s Writing This?”
His inquiry was prompted by a book titled The Future of Truth, authored by Steven Rosenbaum with significant AI assistance. The result, as reviewed by The Times, was a cautionary tale: the book contained more than half a dozen fabricated or misattributed quotes. One such error involved tech journalist Kara Swisher, who famously retorted that the misquote was not only factually incorrect but also suffered from a jarring, uncharacteristic tone.
This incident serves as a focal point for a broader debate: Does AI have a place in professional journalism, or does the outsourcing of “thought-making” fundamentally erode the credibility upon which the entire information ecosystem relies?
A Chronology of the "Hallucination" Crisis
The rise of Generative AI (GenAI) has promised a future of unprecedented efficiency. However, the trajectory of these tools has been marred by a recurring technical phenomenon: the “hallucination.”
- February 2023: Google officially clarified its stance on AI-generated content, emphasizing that its ranking systems prioritize high-quality, original content regardless of how it is produced, provided it adheres to the E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) framework.
- Late 2024–Early 2025: As AI tools became ubiquitous, publishers began experimenting with large-scale automated content production, leading to an influx of low-quality, derivative, and often factually dubious articles.
- May 2026: The publication of The Future of Truth by Steven Rosenbaum brought the risks of AI-assisted research into sharp focus. The book’s reliance on unverified AI outputs led to embarrassing fabrications that damaged its own premise.
- Present Day: Industry leaders like Sam Sifton have begun to publicly reaffirm the necessity of human editorial oversight, distinguishing between using AI for logistics and using it for the core intellectual labor of journalism.
The Mirage of Efficiency: Why AI Hallucinations Matter
Rosenbaum’s defense of his book’s errors—suggesting that the hallucinations “serve as a warning about the risks of AI-assisted research”—is a circular argument that ultimately undermines the author’s authority. In professional journalism, accuracy is not a feature; it is the product.
When an AI “thinks,” it is not engaging in logic, reflection, or the pursuit of truth. It is calculating statistical probabilities of word sequences. Without a human journalist to act as a curator, fact-checker, and skeptic, the output is merely a reflection of the data it was trained on, prone to biases and total fabrications. As Sifton noted, the process of writing is fueled by "adrenaline and fear of errors." It is precisely that fear—the accountability to the reader—that forces a human journalist to verify, corroborate, and iterate. When that fear is removed by automation, the quality of the content inevitably collapses.
Google’s Stance: Quality Over Methodology
A common misconception in the SEO community is that Google’s guidance is a “green light” for AI content. This interpretation is dangerously reductive.
Danny Sullivan and Chris Nelson of Google have consistently maintained that the search engine’s algorithms are agnostic toward the tool used to create content, but they are hyper-vigilant regarding the utility and integrity of that content. Google’s spam policies explicitly state that using automation to generate content for the purpose of manipulating rankings is a violation.
The Content Farm Parallel
Google’s current approach to AI is best understood through the lens of the "content farm" crisis of the early 2010s. When mass-produced, low-quality human content threatened the integrity of search results, Google did not ban human writing. Instead, they introduced sophisticated systems like Panda, the Helpful Content updates, and the E-E-A-T framework.
Today, those same mechanisms are being applied to AI. Google’s algorithms are increasingly capable of identifying the "hollow" nature of AI-generated text—content that lacks original reporting, unique analysis, or the "information gain" that signals genuine expertise. Rosenbaum’s book represents exactly what these systems are designed to discount: content produced without the rigorous editorial accountability that search engines now treat as a primary signal for quality.
The Human Advantage: Accountability as a Ranking Signal
The reason Sifton’s newsletter thrives—and the reason it is the type of content Google’s systems are designed to reward—is not because it is "human-made" in a romantic sense, but because it is "human-made" in an accountable sense.
Accountability is the missing ingredient in the AI-centric content model. When a human writer puts their name on an article, they are staking their reputation on the accuracy of that piece. They are performing the “deep reading” and “question-asking” that AI cannot simulate. For SEO professionals, the lesson is clear: if your content cannot pass the "editor test"—if you would be embarrassed to put your name on it or hand it to an editor—then it is likely destined to be buried by Google’s quality signals.
Implications for the Future of Digital Publishing
Will Sifton’s letter change the industry? Not directly, and not overnight. However, it makes the "human cost" of AI-only content legible to a broad audience.
1. The Death of Commodity Content
As AI makes the production of mediocre content cheaper, the value of that content will trend toward zero. Search engines are already adjusting to prioritize content that offers "originality" that cannot be scraped from the web. The future of visibility lies in reporting, interviews, and synthesis that requires physical presence and intellectual risk.
2. The Return to Expertise
The E-E-A-T framework is no longer just a set of guidelines; it is a business strategy. Publishers who invest in genuine expertise will find their content surfacing more frequently as the web becomes saturated with AI-generated noise. Trust is becoming a scarce resource, and, consequently, a massive competitive advantage.
3. Verification as a Core Competency
In the past, writing was the primary skill of the journalist. In the age of AI, verification is the primary skill. The ability to discern fact from algorithmic hallucination will determine which publications survive the next decade.
Conclusion: The Standards Don’t Move
AI is, by all accounts, an incredible technological leap. It is responsive, adaptive, and improving at an exponential rate. However, the standards that determine whether content earns trust—from readers and from search engines alike—are entirely independent of the technology.
Those standards have been moving in the same direction for as long as the internet has existed: toward higher quality, greater accountability, and the elevation of human insight. Every attempt to "hack" the system by replacing critical thought with automated scale has eventually met the same fate. The algorithms do not yield to the lure of cheap volume; they simply become better at identifying and ignoring it.
As Sam Sifton promised his readers, the "thought-making" of journalism is a human domain. In the pursuit of truth, there is no shortcut, no algorithm, and no substitute for the adrenaline, the fear of errors, and the profound responsibility of an editor who knows that their name—and their reputation—is on the line. The future of truth, despite what the AI-generated books might say, remains a strictly human endeavor.
Pevita Pearce
Lina Hope
