What Google actually did
Google released a spam update on 24 June 2026. It rolled out globally across all languages and finished on 26 June, taking a little over two days. This is the second spam update of the year, following one in March.
Two details stand out. First, Google confirmed this update did not target link spam or the site reputation abuse policy, which are usually the headline targets of a spam update. Second, this followed a busy few weeks for Google in the AI search space:
- On 15 May, Google expanded its spam policies to explicitly cover attempts to manipulate generative AI responses, including AI Overviews and AI Mode. (Google’s spam policies)
- On 3 June, Search Console rolled out a new report showing how often a site’s pages appear in AI Overviews and AI Mode. (Search Console AI performance report announcement)
- On 17 June, site owners were given an opt-out toggle to suppress their pages from appearing in AI Overviews, AI Mode, and Discover.
You can track the exact rollout window for this and any future update on Google’s Search Status Dashboard.
So in the space of about six weeks: a new AI-focused policy, a new AI measurement tool, and then a spam enforcement update. That timing is worth paying attention to, even without a confirmed link between the three.
The bit that got interesting
A handful of site owners with clean records reported traffic drops of 10 to 15 percent after the update, which is unusual for a spam update that is meant to be narrowly targeted. Nobody has confirmed exactly why, but one plausible explanation floating around the SEO community is that Google’s classifiers for detecting mass-produced, templated content are getting more sophisticated, and are occasionally catching more than intended. Google also recently published research on a system for spotting coordinated, templated content at scale (built for video, but using the same underlying logic that could plausibly extend to web content).
The point of these classifiers has never been “is this AI.” Google has been consistent since 2023 that it does not penalise content for being AI-generated. What it penalises is content published in bulk, from templates, with little or no real value to the person reading it, regardless of whether a human or a machine typed it.
What we saw across our own client sites
We pulled Search Console data for a small sample of client sites covering 20 June through 4 July, spanning the update window either side. We’ve anonymised the data below rather than sharing client dashboards directly, since even with names removed, a real screenshot can be identifiable to someone who knows the site.
Across all three, average position held steady through the period, with no sustained shift up or down during or after the 24 to 26 June rollout. There was some day-to-day noise, as there always is. One lower-traffic site showed a sharp single-day click drop, but its ranking position didn’t move alongside it, which points to normal reporting variance on a low-traffic property rather than a real impact. All three sites also dipped slightly on weekends, a completely normal weekly pattern unrelated to the update, and showed a small dip in the most recent day or two of data, which is typically just Search Console’s reporting lag rather than a real drop.
In short: no client site we checked showed a meaningful ranking impact from the June update, which lines up with the pattern above: this appears to be aimed at scaled, low-value content, not well-made pages.
To put that in context, the kind of content this update is understood to target looks like a business publishing the same page fifty times with only the suburb name swapped out (“Plumbers in [Suburb]”), or generating hundreds of blog posts a month from generic prompts with no editing, no original detail, and no real expertise behind them. That’s a very different thing from a handful of well-briefed, well-edited pages.
We do use AI in our content process, but not the way most people picture it. Every piece starts with a call with the client, where we get into the actual detail: history, point of difference, customers, tone, the things that make the business what it is, not just a form to fill in. That conversation becomes the brief we give Claude, written with the same depth we’d give a human copywriter. Once the draft comes back, one of our team goes through it line by line, rewriting anything generic and adding the specific detail an AI can’t know on its own, before it goes anywhere near a client’s site.
That is, as it happens, exactly the distinction Google itself has drawn since its original 2023 guidance on AI-generated content: production method isn’t the issue, quality and genuine usefulness are.
What this means for you
If your site was built properly, with original content specific to your business, this update is very unlikely to have touched you. If you’re working with a provider who is pumping out generic, templated pages to save time, this is the kind of update that starts to catch up with that approach. It’s a good moment to ask what’s actually behind your website’s content, not just whether it’s “AI or not.”
If you’d like us to check how your site is tracking, get in touch for a quick review.
