- Does this Google update really affect me, or is it just for big publishers?
- What exactly did Google change, and what did it say before?
- If I follow the new rules, will traffic return?
- How big is the drop in organic traffic right now?
- What actually changes in the workflow for content creators?
- Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
- What content can't Google replace with an automated answer?
- What concrete risks do those who don't adapt face?
- How to get started: concrete steps to adjust your content strategy now?
- How to tell if the strategy is actually working?
- So, is it worth changing your company's approach now?
- How we see it at SHM
- Frequently Asked Questions
- Google updated its guidelines on October 1, 2026: all AI content must be reviewed by a human editor before publication, including titles, meta descriptions, and image alt text.
- Even those who follow the new rules have no guarantee of traffic recovery: ranking and clicks are two different things, and AI Overviews steal clicks even from well-positioned pages.
- Global organic traffic from Google dropped by 33% between November 2024 and November 2025; with AI Overviews present, only 21% of users click on a link.
- The way forward is to stop producing content in volume and invest in original content with real human oversight: it is the only lever that Google cannot replace with an automated response.
Has your site lost organic traffic in recent months and you're waiting for Google to "fix things"? Google just clearly stated that there is no automatic reset—and that even if you do everything right, the traffic might not come back.
Does this Google update really affect me, or is it just for big publishers?
It concerns anyone who publishes content on the web, even an SME with a company blog or an e-commerce site with automatically generated product descriptions. If you use AI tools to produce text, titles, or descriptions, you are directly involved.
October 1st, 2026, Google has updated its guidelines on the use of generative artificial intelligence in content production. The main update: every piece of content produced with AI must be checked and verified by a real person before going online. Not just the text, but also titles, meta descriptions (the short description that appears in search results), structured data (the structured data that helps Google understand the content), and image alt text.
This isn't just a vague recommendation. Google has made it clear that publishing large volumes of AI pages without adding real value violates its policies against so-called scaled content abuse — that is, the abuse of content produced at scale. The rule applies to a manufacturing company that generates 200 product sheets with ChatGPT exactly as it does to a large online publisher.
| Your situation | Are you involved? | why |
|---|---|---|
| You use AI to write blog articles | Yes | A human editor is needed for every piece before publication |
| You generate meta descriptions and titles with AI | Yes | Google explicitly includes these elements in the new guide |
| You have automatically written product pages | Yes | They fall under the 'scaled content abuse' risk if you don't add real value |
| Write everything by hand without AI | No, for now | You aren't directly impacted, but the competitive landscape is changing anyway |
| You already use AI as support with human review active | No, for now | You are already aligned with Google's new guidelines |
What exactly did Google change, and what did it say before?
Google has made the requirement for manual fact-checking—meaning verification by a person—explicit for any AI content, extending the check to all elements on the page, not just the body text. Previously, this requirement was not written in black and white.
Google has been working on content guidelines for years. The journey went through Panda (2011), then E-E-A-T — an acronym for Experience, Expertise, Authoritativeness, Trustworthiness — and the Helpful Content updates. As we explain in detail in the article on Google and mandatory fact-checking for AI content , the October 2026 update is the first time that the requirement for human review is explicitly written and applied to every single element of the page.
The practical difference compared to the past: before, you could interpret the guide as a generic invitation to quality. Now Google says there must be a real editor — a person with enough context and authority to improve or block publication — for every piece of content that goes out.
| Page element | Mandatory human review (before) | Mandatory human review (after) |
|---|---|---|
| Body text | Implicit | Explicit |
| Title (title tag) | Not mentioned | Explicit |
| Meta description | Not mentioned | Explicit |
| Structured data | Not mentioned | Explicit |
| Image alternative text (alt text) | Not mentioned | Explicit |
It is worth noting that Google's spam policies — the rules against junk content — now officially also cover AI Overviews and AI Mode . This means that control doesn't stop at classic results: it also concerns how Google builds its automatic answers.
If I follow the new rules, will traffic return?
Probably not, at least not automatically. Google says it clearly: improving content can take time and doesn't guarantee a visible traffic recovery. Ranking and clicks are two different things.
This is the point that interests those who have seen their visits drop in recent months the most. Google's answer is uncomfortable: even if the site improves and Google re-evaluates it positively, that page competes in a completely different search environment than the one that generated traffic before.
AI Overviews—the automatic response blocks that Google shows at the top of results—take up space, answer user questions, and reduce the need to click on any links. As the analysis of the phenomenon shows Google Zero and the disappearance of organic traffic , a page can climb positions without gaining a single click.
The data is clear: when an AI Overview is present, only 21% of users click on an external link, compared to 56% when there is no automated response. The decline is structural , not temporary. Even an article perfectly rewritten and approved by Google cannot recover the clicks that Google has decided not to send anymore.
How big is the drop in organic traffic right now?
Organic traffic from Google dropped by 33% globally between November 2024 and November 2025. In the United States, the decline was 38%. This is not a temporary adjustment.
The numbers collected from thousands of editorial publications confirm a structural trend. According to Reuters Institute data Out of 268 editorial managers, organic traffic from Google has dropped by 33% globally in one year. This phenomenon doesn't just affect those who publish news: any site that relies on organic search to bring in visitors is exposed.
Google's AI Overviews were present in 10% of desktop informational searches in March 2025. By September 2025, they were already at 30%. Growth is rapid and shows no signs of slowing down.
However, there is one piece of data that goes in the opposite direction: news content recorded a +103% growth on Google Discover between March 2025 and March 2026. Discover — the personalized content feed that Google shows on smartphones — is becoming an alternative channel for those who produce original and updated content.
What actually changes in the workflow for content creators?
The publishing process must include a real human review phase, not just a quick read-through. This applies to every element on the page, not just the main text.
- 1 · Editor / AI Content production The text is written or assisted by AI. This step doesn't change.
- 2 · Human editor Fact-checking A person with subject matter expertise checks every claim, data point, and source cited in the text.
- 3 · Human editor Technical element review Title, meta description, structured data, and image alt text are reviewed and approved by a person, not left to automatic output.
- 4 · Editorial Manager Final decision: publish or block Whoever has editorial authority decides if the content is ready. If it doesn't add real value compared to what already exists, it doesn't go online.
- 5 · SEO / Analytics Team Post-publication monitoring Rankings, impressions, and clicks are tracked separately, because an improvement in ranking does not equate to a recovery in traffic.
If your current workflow is: generate 50 articles with AI → quick review → publish, adding a better prompt doesn't solve the problem. Google isn't asking for more polished text on the surface: it's asking for a real editorial process. How does it deepen the analysis on Google and anti-spam rules on AI in Search , the distinction between helpful content and scaled content is now at the heart of the main ranking systems, not a separate filter.
Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
The impact varies based on how much the site depends on organic traffic and how much content is produced with AI. Here are three typical situations with priority actions.
Illustrative examples: the numbers and situations are meant to show the reasoning, they are not real customer data.
| Company type | Typical situation | Main risk | Priority action |
|---|---|---|---|
| E-commerce (e.g., technical clothing) | 500+ product pages with automatically generated descriptions, AI meta descriptions | Scaled content abuse: similar pages without real added value | Audit of product pages: which ones add unique information? Human review for main categories |
| Professional firm (e.g. accountant) | Blog with tax articles produced with AI, 2-3 pieces per week | Generic content that AI Overviews already cover without needing a click | Switching to content with original data, practical cases, direct team experiences |
| Manufacturing company (e.g., components) | Site with technical data sheets and industry articles, B2B traffic from organic search | Drop in visibility on informational queries where AI Overviews answer directly | Invest in deep technical content that AI cannot synthesize without losing precision |
What content can't Google replace with an automated answer?
Content that includes original data, direct tests, case studies, and first-hand experience is what AI Overviews cannot replicate without losing value. It is also what Google is most interested in showing.
Google has less reason to reward content that repeats the same information available on fifty other pages. If a question has a standard answer, the AI Overview provides it directly and no one clicks. But if your content contains something that doesn't exist elsewhere — an internal analysis, a comparison based on real data, a documented direct experience — then it becomes difficult to replace.
According to an analysis by Graphite on 900,000 pages published in April 2025 , 74.2% of new web content already contains AI to a partial or total extent. Only 2.5% is entirely AI-generated. The problem isn't AI itself: it's that when everyone uses the same tools to produce the same content, the average value drops and Google has no reason to send traffic to any of them.
| Content type | Replaceable by AI Overview? | Residual SEO value |
|---|---|---|
| Generic definitions and explanations | Yes | Low: the automated response already covers the question |
| Standard step-by-step guides | Yes | Medium-low: depends on depth |
| Original data and internal research | No, for now | High: not replicable without direct access |
| Case studies with real numbers | No, for now | High: documented direct experience |
| Technical content with precise specifications | No, for now | High: the automated summary loses precision |
| Opinions and analysis signed by experts | No, for now | Medium-high: depends on perceived authority |
What concrete risks do those who don't adapt face?
The main risk is not an immediate penalty, but a gradual slide into invisibility: content that is no longer shown, neither in classic results nor in Google's automatic answers.
Google has integrated content quality signals into its core ranking systems. There is no longer a single update to wait for to see if changes have worked. The system evaluates continuously, and those who produce high-volume content without real added value accumulate a quality debt that is difficult to recover from quickly.
As the analysis on explains what to do right now for visibility before the next Google update , waiting for the problem to show up in Analytics data is already too late: at that point, the damage is done and recovery takes months.
There is also a less obvious risk: unreviewed AI content can contain factual errors. If Google detects them — or if users report them — the domain's authority suffers, and this impacts all the pages on the site, not just the problematic ones.
| Risky behavior | Probable consequence | Timing |
|---|---|---|
| Publishing hundreds of AI pages without review | Scaled content abuse: penalty or deindexing | Variable, can be fast |
| Meta descriptions and titles generated automatically without review | Reduction in perceived relevance, drop in CTR | Gradual |
| Generic content that replicates information already available | Invisibility in queries covered by AI Overview | Ongoing |
| No separate monitoring of ranking vs. clicks | Wrong decisions based on incomplete data | Immediate |
How to get started: concrete steps to adjust your content strategy now?
The starting point is an audit — that is, a systematic review — of existing content, not the production of new content. First you understand what is there, then you decide what to keep, improve, or eliminate.
The temptation is to add a review step to the current workflow and consider the problem solved. It doesn't work like that. The problem is structural: if the production model is based on volume, adding a check at the end doesn't change the average quality of the content.
- 1 · Marketing / SEO Manager Content catalog audit Map all site pages by content type, traffic received in the last 12 months, and presence of AI Overviews in target queries. Identify pages that no longer receive clicks despite their ranking.
- 2 · Editorial team Classification: keep, improve, delete For every page: does it have original data or direct experience? Does it add something that cannot be found elsewhere? If not, consider whether to improve it significantly or remove it so as not to penalize the domain.
- 3 · Designated human editor Define who is responsible for the review Appoint an editor with subject matter expertise for each content area. It can't be someone who doesn't know the industry: they must have enough context to judge if information is accurate.
- 4 · SEO team Separate ranking metrics from traffic metrics Monitor impressions, average position, and clicks as three distinct metrics. An improvement in ranking without growth in clicks signals the presence of AI Overviews on the query.
- 5 · Marketing manager Identify alternative channels Google Discover, newsletters, social media, and citations in AI Overviews are growing channels while classic organic traffic is declining. As the analysis on how to appear in Google AI Overview responses shows, visibility is built on multiple fronts.
How to tell if the strategy is actually working?
Looking at ranking isn't enough. You need at least three separate metrics: position, impressions, and clicks. Only then can you understand if the problem is content quality or the structure of the search results.
As the analysis on explains SEO and site visibility in the AI era , the very concept of SEO success needs to be redefined. A site can improve content quality, climb the rankings, and still see clicks drop — because the AI Overview answers the question before the user even reaches the site. This isn't a strategy failure: it's the new context we're operating in.
| Indicator | What it measures | Positive signal | Red flag |
|---|---|---|---|
| Average position (Google Search Console) | Where the site appears in results | Constant improvement on target queries | Stable or declining after content review |
| Impressions | How many times the site appears in results | Growth on new queries | Drop in historical queries (AI Overview covers them) |
| CTR (Click-Through Rate) | % of users who click on the link | Above 3-4% on queries without AI Overview | Under 2% on queries with AI Overview active |
| Discover traffic | Visits from the personalized Google feed | Growth on original and updated content | Absent: the site does not produce fresh content |
| Citations in AI Overviews | The site is cited in automated responses | Presence as a source in AI answers | Total absence even on industry queries |
So, is it worth changing your company's approach now?
It is worth it if your site has lost organic traffic in the last 12 months, you produce content with AI without a structured review process, or you have pages in volume with little added value. Waiting won't recover traffic: taking action at least improves quality and reduces the risk of penalties.
A middle ground Your site has stable traffic but you're noticing a drop in CTR (meaning fewer people are clicking even though the site appears in the results). It's worth doing an audit to figure out which queries AI Overviews are stealing clicks from, then deciding where to invest in original content.
Not yet if Your site already produces content with original data, real human review, and doesn't depend on AI page volumes. In this case, you're already aligned with the new guidelines and can focus on Discover and citations in AI Overviews as growth channels.
How we see it at SHM
For us, AI is a tool to free up time—not to fill pages. When a company uses AI to produce 300 articles a month without a real editorial voice, it isn't saving time: it's accumulating a quality debt that Google will eventually come to collect.
What Google is asking for with this update is exactly what we have always believed: technology does the repetitive work, but the decision on what is worth publishing remains human. An editor who knows the industry, who knows when information is wrong or when content adds nothing — that cannot be automated.
In projects where we support companies on SEO for AI and on Strategic AI integration , the first step is always an audit: understanding what is already there, what works, and what needs to be eliminated. Only then can you build a sustainable production workflow. If you want to understand where to start for your company, our team is available for a initial consultation .
You can dive deeper into all the topics related to this scenario in our collection dedicated to SEO and visibility for AI .
Frequently Asked Questions
Not necessarily. The problem isn't AI itself, but the lack of real added value. Google isn't asking to get rid of AI: it's asking for every piece of content to be reviewed by a competent person before publication. Lost traffic also depends on the presence of AI Overviews, which take away clicks regardless of content quality.
Google itself warns that improvements can take time and there's no guarantee of a visible recovery. Generally, we're talking about weeks or months for re-evaluations. The most important thing is not to expect a quick reset and to monitor the right metrics: rankings, impressions, and clicks separately.
Yes. If product descriptions are automatically generated with AI and do not add original information compared to the manufacturer's, they fall under the risk of scaled content abuse. Human review must also cover meta descriptions and image alt text.
Check Google Search Console: if impressions (how many times you appear in results) remain stable or grow but clicks drop, it's likely that an AI Overview is answering the question before the user reaches your site. A CTR below 2-3% on informational queries is a typical sign.
Google talks about a person with enough context, access, and authority to improve content or block its publication. A quick read isn't enough: the editor must be able to fact-check, assess whether the content adds real value, and approve every element on the page, including the title, meta description, and alt text.
Yes, especially for those who produce original and up-to-date content. News content saw a +103% growth on Discover between March 2025 and March 2026. Discover rewards freshness and relevance to the user, not just ranking for specific queries. It is a channel to manage in parallel with classic SEO.
Traffic doesn't come back by waiting: it comes back—partially—only to those who stop producing content for Google and start producing it for the people Google still can't replace.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
