Recently, Sam Sifton, who hosts The Morning news bulletin for The New York Timespublished a letter to its readers with an unusual subject: “Who is writing this?»
His impetus was a new book called “The Future of Truth,” written by Steven Rosenbaum with significant help from AI. The times reviewed the book and found more than half a dozen misattributed or entirely fabricated quotes conjured up by the AI, including one attributed to tech journalist Kara Swisher. Swisher’s response said not only was the quote false, but that “I also feel like I have a stick up my butt.”
Rosenbaum’s defense that hallucinations “serve as a warning about the risks of AI-assisted research and verification” is the kind of phrase that would be more compelling if it appeared in another book.
Sifton took the opportunity to tell his readers something he clearly felt they deserved to hear directly. The Morning is built by humans, for humans. His team can use AI to find information verified elsewhere. They can use it for editorial logistics, buying time for more reporting, but the thinking, questioning, close reading and writing that follows – these are tasks done by journalists without a chip. “I write fueled by adrenaline and the fear of mistakes,” he told his readers. “And I promise you that will never change.”
What Google Advice Really Says
In February 2023, Danny Sullivan and Chris Nelson published the Google report tips on AI-generated content. This position, which has not changed much since and was further reinforced recently in Matt Southern’s report on Google’s new AI research guideis it: Google’s ranking systems aim to reward original, high-quality content which demonstrates EEAT (expertise, experience, authority and reliability). The focus is on the quality of the content and not how it is produced.
This looks, after a quick read, like a green light for AI content. This is not the case, or at least it is not a green light without extremely important conditions.
Google’s guidelines specifically state that using automation to generate content with the primary purpose of manipulating search rankings violates its anti-spam policies. And he draws an analogy that SEO professionals should analyze and evaluate: About a decade before the 2023 guidelines were written, understandable concerns existed regarding content farmswhich mass-produced large volumes of human-generated content. No one thought it reasonable to ban all human-generated content. Instead, Google has improved its systems to reward quality. The Useful Content System, the EEAT Framework, the Information Gain Patent, the ongoing updates of the Quality Assessor Guidelines until 2025 – all this is the same enforcement mechanism, applied again, with greater sophistication.
Rosenbaum’s book is exactly the type of content that Google’s systems are designed to identify and discard. Not because he used AI, but because he used AI negligently, without the verification, original reporting and editorial accountability that comes with it. Google quality signals are trained to detect.
Sifton’s newsletter is exactly the kind of content these same systems are designed to reward. Not because it is man-generated, but because it is produced by people with real expertise, direct experience and accountability to a specific audience. It’s built by humans, for humans, in precisely the way Google’s helpful content tips always intended.
Will Sifton’s letter change anything?
The question at the center of this commentary is whether Sifton’s take on the growing role of AI will change what Google does, change the way practitioners write for AI, or change the way they gain visibility into AI.
The honest answer is no, not directly, and that’s the point.
Google’s advice has been consistent since February 2023. It was consistent before that in spirit, through Panda in 2011, through EAT, through the Useful Content Update in 2022, until the transition to EEAT later in the year. What changes is only the acuity with which people spot it on the horizon.
What Sifton’s letter does, what Google’s technical documentation cannot do, is make the human cost of the alternative legible. Rosenbaum’s Kara Swisher hallucination is not a borderline case or a technical failure. This is what happens when thinking is entirely outsourced, when questioning stops, when no one writes fueled by adrenaline and fear of mistakes. This is a book about the future of truth that cannot be trusted.
For SEO professionals, the practical implication hasn’t changed since Amit Singhal’s 23 Panda Questions in 2011. Does the article provide original content or information, original report, original research or original analysis? Does it have the kind of quality you’d expect to see referenced in a magazine, encyclopedia, or book? Would you be comfortable giving this to your editor and putting your name on it?
Sifton’s promise to his readers is that he would. This responsibility is not a stylistic choice. This is the entire mechanism by which trust is established with an audience and by which Google’s systems learn to surface content worth publishing.
The real lesson
AI is not indifferent. It is responsive, adaptive, and improving faster than any previous technology transition in the history of the industry. That’s exactly what makes it useful and that’s exactly what makes the question of how to use it so important.
But the standards that determine whether content is worthy of trust, both from readers and Google’s ranking systems, don’t evolve on AI’s timetable. They have been going in the same direction since Google existed. Every approach that assumed these standards would give way to scale, automation, and the next optimization trick ended up with the same thing.
They don’t give in. They move on as if nothing had happened.
More resources:
Featured image: Roman Samborskyi/Shutterstock
