For companies producing content with AI support, an announcement like this sounds like a red alert. The good news first: Google has nothing against AI-generated content. The less good news: it matters a great deal how and for what you use it.
1. What started on 18 August 2026
The Google Search Central account announced the rollout of the August 2026 spam update. Google describes it as a regular update to improve search result quality — routine, not a special case. The rollout was expected to complete within a few days and applies globally, across all search languages.
You can check the current state at any time in the Google Search Status Dashboard, where Google documents every ongoing search event with start and end dates. It is the only reliable source for whether an update is still running or already finished.
- 18 August — the August 2026 spam update starts, worldwide, all search languages.
- 2024 — the year Google began acting systematically against scaled content abuse.
- 14 June 2026 — the date from which Google classifies back-button hijacking as spam.
Put briefly: solid SEO work that follows Google's guidelines has nothing to fear from a spam update. It gets uncomfortable where content is made primarily for the algorithm rather than for people — and AI at scale produces exactly that pattern with unusual ease.
2. Spam update or core update? The difference decides your response
Many site owners lump both update types together and derive the wrong measures from it. A core update adjusts how Google assesses relevance and quality in general. A spam update targets concrete breaches of the spam policies.
The practical difference: after a core update the answer is "improve the content". After a spam update it is "find the breach and stop it". Producing more content after a spam update can make the problem worse.
Worth knowing: Google's spam detection systems run continuously, not only on update days. Announced updates simply mark larger, broader adjustments — Google communicates them so site owners can interpret unusual swings in their figures.
3. The first days: little movement, plenty of unease
At the start of the rollout the common volatility trackers showed no unusual swings: neither the SISTRIX update radar nor the Semrush volatility sensor reported abnormal movement for the German index. In its first days, the August 2026 spam update did not move rankings much.
That is typical, and it is not a reason to relax by inversion. Spam updates need several days to roll out fully, and their effects concentrate by nature on a small part of the index — the pages that actually breach guidelines. A quiet sensor chart mostly means: the vast majority of sites are unaffected.
For your own analysis that means looking at your own numbers in Google Search Console rather than at industry sensors. Compare impressions and clicks in the week after the update with the preceding weeks — segmented by page type, not as a single total.
4. AI content is not the problem — scaled content abuse is
This is where the biggest misunderstanding sits. Google does not forbid AI-generated content. The company has stressed for years that generative AI can be useful in researching and structuring content. Quality and purpose are what count.
What Google acts against is called scaled content abuse: producing pages in large numbers whose main purpose is to manipulate rankings rather than help users. The policy has applied since 2024 and was deliberately worded to bite even when it cannot be proven that low-value content was produced purely automatically.
The decisive sentence in Google's position: automation — generative AI included — counts as spam when its main purpose is manipulating search results. The criterion is not the tool but the intent and the outcome.
The situation is sharpened by the fact that assistants such as ChatGPT, Gemini and Claude have made producing text, and whole page inventories, extremely cheap. Quickly assembled sites without editorial substance fall into the same pattern. A clean AI-supported editorial process with subject-matter review, source work and clear authorship is a long way from that category.
5. Expired domain abuse, site reputation abuse and link spam
Alongside scaled content abuse, Google pursues two further patterns under its more recent policies.
Expired domain abuse describes buying lapsed domains to recycle their accumulated authority for entirely unrelated content. Buying a former club domain and turning it into an affiliate portal is exactly this.
Site reputation abuse — often called parasite SEO — means placing third-party content on an established domain to skim its trust. The classic case: a news portal renting out subdirectories to voucher or casino providers.
Add the perennials: content created and distributed in a deceptive context, misleading redirects, and link spam. For mid-sized companies these points are usually uncritical — it gets critical where an agency advertises "fast rankings" or link packages.
6. Back-button hijacking: officially spam since June 2026
A rule change many missed: since 14 June 2026 Google classifies back-button hijacking as spam without qualification. The term covers manipulating the browser's back button so that a click leads not to the previously visited page but to a third-party site, unwanted recommendations, or straight into advertising.
Google's reasoning is consistently user-centred: back-button hijacking interferes with browser functionality, breaks the expected journey and causes frustration. Users report feeling manipulated and subsequently visiting unfamiliar sites less often. Because the practice has increased, the guidelines were adjusted.
This is not an exotic edge case. Some consent layers, exit-intent overlays, pop-up plugins and aggressive ad scripts interfere with browser history without the site owner knowing. Testing costs two minutes — see the self-check below.
7. An EU special case
For the European market there is an additional wrinkle. According to Reuters, the European Commission holds a document in which Google signals an adjustment of its site-reputation-abuse rules within the EU — apparently to head off a threatened penalty. What that adjustment looks like in detail has not been published.
It is conceivable that parasite SEO would become permissible again in the EU. Google itself continues to stress that it wants to protect users from deceptive practices. There is no official change so far; as of today the existing rules apply.
Our recommendation is unambiguous all the same: do not build a visibility strategy on a rule that is currently being negotiated politically. Investing in borrowed domain authority today may leave you with a worthless asset tomorrow — and in AI answer systems something else counts anyway, namely demonstrable authority of your own.
8. Are you affected? A five-step self-check
- Evaluate Search Console by segment. Compare impressions and clicks from 18 August with the preceding weeks, separated by page type (home, service pages, blog, categories). A drop that hits only one page type is a strong hint of a structural problem.
- Check manual actions. Open the "Manual actions" report in Search Console. An entry there means a manual sanction rather than an algorithmic effect — and after fixing it you have to submit a reconsideration request explicitly.
- Inventory your content honestly. List every page created in large numbers in a short time — location, keyword or category variants built to a formula. Ask of each: what independent value does it offer? Without a good answer it belongs merged or deleted, not optimised.
- Test the back button. Reach your site from a search result, click two levels deep and press back. If you do not land where you expect, check consent layers, pop-up and ad scripts. This applies especially on mobile.
- Look through external providers and subdirectories. Check whether third parties publish on your domain (guest posts, rented directories, affiliate partners) and where bought links come from. Third-party content on your domain is your risk, not your partner's.
9. The update punishes intent, not tools
For the vast majority of companies the August 2026 spam update is not an event that forces immediate action. Volatility stayed low at the start, and Google stresses that sites following the guidelines and delivering relevant content usually see no change.
The real value of the update lies in the reminder it carries. The dividing line does not run between "written by humans" and "written by AI", but between "made for people" and "made for the algorithm". Use AI as a tool for faster, better-researched, editorially reviewed content and you gain. Use it as a volume machine and you build a risk that comes due sooner or later.
Our assessment: the best protection against any future spam update is unspectacular — fewer pages but better ones; recognisable authorship and subject expertise; clean technology without tricks. That same substance decides in parallel whether AI answer systems cite you as a source. One effort, two effects.
10. FAQ: the key questions about the August 2026 spam update
Does Google now penalise AI-generated content?
No. Google has repeatedly made clear that generative AI is legitimate for researching and structuring content. Quality and purpose decide: automation counts as spam when its main purpose is manipulating search results. A carefully researched, editorially reviewed article stays uncritical even if AI helped write it.
What exactly is scaled content abuse?
Scaled content abuse describes producing pages in large numbers, often with low-value content, whose main purpose is to increase the chance of prominent placement in search. The policy has applied since 2024 and was framed deliberately so that it bites even where it cannot be proven that the content was produced purely automatically.
How do I tell whether the spam update hit me?
Compare impressions and clicks in Google Search Console from 18 August 2026 with the preceding weeks, segmented by page type. Also check the "Manual actions" report. If you see losses, measure your site against Google's spam policies — according to Google, noticeable ranking losses after a spam update are typically the consequence of policy breaches.
How long does recovery after a spam update take?
Google describes recovery as something that can stretch over months: only once the automated systems establish over time that a site complies with the spam policies can that have a positive effect on ranking. Do not expect a quick return to old positions — what is required is sustained clean-up rather than short-term correction.
