Google’s Values Don’t Change But AI Makes It Harder To Ignore Them
Recently, Sam Sifton, editor of the Morning newspaper The New York Timespublished a letter to its readers with the unusual title, “Who Writes This?”
His prompt 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 a half-dozen uncredited or entirely fictional quotes put together by AI, including one attributed to tech journalist Kara Swisher. Swisher’s response was that not only was the quote wrong, but “I feel like I have a stick up my sleeve.”
Rosenbaum’s defense that the hallucinations “serve as a warning about the dangers of AI-assisted research and validation” is the kind of sentence that would be more convincing if it came from a different book.
Sifton used this time to tell his students what he felt they needed to hear firsthand. Tomorrow is built by people, people. His team can use AI to find information that can be verified elsewhere. They may use it to organize things, to buy more reporting time, but thinking, asking questions, reading deeply, and writing the following – those are the tasks that journalists do without chips. “I write motivated by adrenaline and fear of mistakes,” he told his students. “And I promise you that will never change.”
What Google’s Guidelines Actually Say
In February 2023, Danny Sullivan and Chris Nelson published Google’s guide to AI-generated content. The position, which hasn’t changed significantly since then and was reinforced recently in Matt Southern’s reporting on Google’s new AI search guidelines, is this: Google’s ranking systems aim to reward original, high-quality content that demonstrates EEAT (expertise, knowledge, authority, and credibility). The focus is on the quality of the content, not how it is produced.
That sounds, on quick reading, like a green light for AI content. It’s not, or at least it’s not a green light except in very important circumstances.
Google’s guidelines specifically state that using automation to generate content for the primary purpose of manipulating search rankings violates its spam policies. And it draws an analogy that SEO experts should analyze and evaluate: almost ten years before the writing of the 2023 directive, there was an understandable concern about content farms, which produced a lot of human-generated content. No one thought it reasonable to ban all human-generated content. Instead, Google has developed its own systems to reward quality. The useful content plan, the EEAT framework, the copyright to gain knowledge, the continuous Quality Rater guidelines are updated until 2025 – it’s all the same enforcement method, used again, with great skill.
Rosenbaum’s book is exactly the type of content Google’s systems are designed to identify and discount. Not because it uses AI, but because it uses AI carelessly, without validation, original reporting, and accountability for programming trained to detect Google’s quality signals.
Sifton’s newsletter is exactly the type of content that similar programs are designed to reward. Not because it is produced by people, but because it is produced by people who have real knowledge, direct experience, and accountability to a specific audience. It’s built for the people, by the people, in the way that Google’s always-intended useful content direction is.
Will Sifton’s Book Change Anything?
The question at the center of this analysis is whether Sifton’s view of the growing role of AI will change what Google does, change the way experts write about AI, or change the way they win in AI visibility.
The honest answer is no, not exactly, and that’s the point.
Google’s guidance has not changed since February 2023. It was unchanged before that in the air, with Panda in 2011, with EAT, with a useful content update in 2022, with the transition to EEAT later that year. What changes is how smart people see it on the horizon.
What Sifton’s book does, Google’s technical documentation can’t, is make the human costs of alternatives readable. Rosenbaum’s Kara Swisher hallucination is not a peripheral or technical failure. What happens when thinking is completely removed, when questioning stops, when no one writes motivated by adrenaline and fear of mistakes. It is a book about the uncertain future of truth.
For SEO experts, what works has not changed since Amit Singhal’s 23 Panda questions in 2011. Does the article provide original content or information, original reporting, original research, or original analysis? Does it have the kind of quality you would expect to see referenced in a magazine, encyclopedia, or book? Would you feel free to give this to your editor and put your name on it?
Sifton’s promise to his students was that he would. That accountability is not a style choice. It’s how credibility is built with viewers, and how Google’s systems learn to display content that’s worth displaying.
A Real Lesson
The AI doesn’t care. It is responsive, flexible, and evolving faster than any previous technology change in the history of the industry. That’s what makes it useful and that’s what makes the question of how to use it so important.
But the standards that determine whether content earns trust, for readers and Google’s ranking systems alike, don’t come from an AI system. They’ve been going the same way for as long as Google has been around. Every approach that thought those standards would provide scale, automation, and the next best practice found the same thing.
They don’t give up. He leaves as if nothing happened.
Additional resources:
Featured image: Roman Samborskii/Shutterstock
