by Jed C Jones, Founder, Ambient Array
I’ve been “doing SEO” as an agency for untold numbers of clients, cumulatively, since 2007. It’s been fascinating to see the progression in what Google (and other top search engines) has viewed as rank-worthy over the years.
Grossly oversimplifying things, what Google has valued over the years has evolved something like this:
More keywords (quantity/stuffing) → Better keywords (quality/being on-point), semantic search and meta tags → More inbound links (quantity) → Better inbound links (quality/PageRank) → Aligned content pillars and stickiness → Simply making good content for a defined audience
In other words, it has approached, over time, the standard of “simply serve the right audience with good content and make it clear how you can materially help make their lives better.” Google has been rewarding this standard increasingly over the years. And that makes sense: gaming the system has never been smiled upon by Google (to the chagrin of SEOs everywhere).
Around 2022, AI and LLMs started getting good enough to make it WAY easier for content creators to produce content automatically. So easy, in fact, that we are all very familiar now with the term “AI slop”: factory-generated content that metaphorically tastes like mushy, boiled cauliflower with no salt or spices added.
And, of course, search has caught on. Now, Google is updating the rules again. Currently, if you want to be found online for free, you need to not only make good content, you need to make content that is original, accurate, shows effort and exhibits talent or skill to write.
Google rewrote its helpful content guide on October 1st. If you’re a CMO running any content program on AI-assisted production, this one deserves your full attention. Google still doesn’t penalize AI content outright. What changed is that it has now spelled out, clearly and publicly, what it considers low-effort, and a lot of what passes for “reviewed AI content” today is going to fail that test.
Here’s what changed and what it means for the decisions in front of you right now.
Why this update is different
Google updates its guidelines regularly, but there’s a pattern worth knowing. In 2011, a Google spokesperson publicly flagged content farms as a problem. Panda dropped the next month. In 2012, over-optimization was flagged publicly. Penguin followed. This past May, Google clarified that spam policies apply to AI Overviews. A core update came six days later.
The October 1st rewrite looks like another signal flare. Google has also run four spam updates this year compared to one in all of 2025, and the gaps between them keep shrinking. Google’s Gary Illyes said publicly that scaled content is now a bigger problem than link spam. Coming from Google, that is not a throwaway line.
If you’re waiting for a penalty before you act, you have the sequence backwards.
Want to keep getting ranked in organic search while improving your GEO/AEO rankings? Check out these four actionable tips:
The four tests Google now applies
Google updated its quality rater guidelines to judge content on four dimensions. Quality raters are real people Google pays to score search results. They don’t rank pages directly, but their scores train the systems that do. So when Google tells raters what to look for, it is telling the ranking systems what to reward.
Effort. Google defines this as how much human work went into creating the content. High-effort examples include original analysis and firsthand demonstrations. Low-effort examples include pages generated automatically from feeds and AI-produced text without meaningful human oversight. Citing sources does not substitute for original effort; Google says that explicitly.
Originality. Does the content offer something that isn’t already available on other sites? A remix of the top ten results fails. Your own data, your own tests, your own point of view pass. A proprietary dataset your team built, an insight from a customer call, a test result you ran internally: a language model simply cannot produce that material on its own.
Talent or skill. Google’s examples include clear writing, well-produced video, and embedded tools that actually solve a problem. Interestingly, Google also says not everything needs a credentialed expert. Real experience, communicated clearly, is enough. What it wants is evidence that a capable human shaped the output.
Accuracy. For content touching health, money, or safety, the bar is especially high. Every fact needs to be sourced and checked against the original. Google now calls this review “critical,” and it specifically includes titles, meta descriptions, and image alt text.
What’s getting punished
The patterns Google has been hitting this year:
- Scaled pages. Hundreds of thousands of pages published quickly, often on topics the site has no real expertise in. Templated city pages and product comparison pages with a few words swapped between them fall into this category.
- Fake authors. AI-generated headshots, invented names, fabricated credentials. Google now calls this deception, in writing.
- Self-promotional listicles. “Best X for Y” content published by the brand that makes X, with that brand conveniently at the top. According to research Lily Ray published in September, Google’s AI Overviews now cite 38% fewer of these lists and leave the publishing brand out of the recommendation 83% of the time.
- Paid placements in third-party lists. There’s a growing market selling spots in “best of” lists, often priced between $135 and $250, to influence AI answers. If those spots include links that pass ranking credit, Google treats it as link buying, which it has penalized for years.
How to use AI without creating the problem
None of this means stop using AI. It means use it differently.
Start with something the model doesn’t have. A customer call transcript, proprietary data your team collected, a strong contrarian opinion, a test you ran. Feed that material in and let the AI help you shape the argument, stress-test it, and find the gaps. Let it do the structural work. You do the thinking.
Fact-check everything by hand. Open every link. Check every number and claim against the original source. Do not delegate verification to the AI. Google specifically flagged this step as critical.
Read every word of the output before publishing. If it hedges constantly, has no real opinion, or circles a point without landing on it, rewrite it until it sounds like a person with a view.
Put a real author’s name on it with a real, clickable bio. And if AI did substantial work on the piece, say so. Google doesn’t penalize AI use. It penalizes low-effort AI use.
Don’t let AI scale you past your actual expertise. That’s where the real risk lives. A model will happily scale intuition dressed up as analysis, and at volume an ungrounded claim doesn’t stay small; it shows up on every page the template touches. Producing content on topics your team has no genuine depth in, letting the model run ahead of your understanding, and treating “accept all changes” as an editorial process is exactly the pattern Google is describing.
So the bar is clear: effort, originality, skill, and accuracy. If your content program can show all four, you’re in a defensible position. If it can’t, the update Google appears to be signaling will find it.
Find out how to grow your marketing with the professional marketing minds at Ambient Array.
Original source: Google Just Defined “Low-Effort” Content (Big Deal for SEO)
