In May 2026, Google updated its spam policy to explicitly include attempts to manipulate AI systems integrated into search — particularly AI Overviews and AI Mode. Therefore, techniques such as artfully constructed “best-of” listicles or so-called Recommendation poisoning are now officially included in prohibited practices.
Furthermore, the update broadens the very definition of spam: it's no longer just about deceiving human users, but also about manipulating the generative models that comprise Search responses. Consequently, content designed to «poison» AI responses can be directly penalized, regardless of traditional organic ranking.
In summary, for Italian SMEs, this means reviewing their content marketing and SEO strategy with stricter criteria. We at SHM Studio We recommend an immediate audit of existing content, with particular attention to comparative pages and product recommendation sections. Those who wish for an operational comparison can contact the team at SHM Studio for a personalized evaluation.
Google's new definition of spam
On May 15, 2026, Google published a revision to its spam policy. The update was reported by The Verge and was reported by Search Engine Land. The substantial novelty concerns the extension of the concept of spam to artificial intelligence systems integrated into search.
The new wording explicitly states: spam includes attempts to manipulate Google Search generative systems, such as AI Overviews and AI Mode. Therefore, it is no longer sufficient to avoid traditional keyword stuffing or link farming practices. Now, consideration must also be given to the impact of content on the LLM models that comprise automated responses.
This passage is significant. In fact, until now, the spam policy was primarily geared towards protecting human users. With this update, Google is also protecting the integrity of its AI system.
What is meant by AI response manipulation
Among the techniques now explicitly prohibited are two main categories. The first is the production of «best-of» listicles constructed with the aim of influencing AI-generated recommendations. The second is the Recommendation poisoning, which is the injection of distorted content into the textual corpora that LLM models use as a source.
In addition, Google flags attempts to insert hidden instructions into texts—a practice analogous to prompt injection — to guide AI Overview responses toward specific products or brands. These techniques, according to Google Search official documentation, will be treated with the same measures applied to traditional spam.
In particular, the sites involved risk demotion in organic search results and exclusion from AI-generated answers. Thus, the potential impact is double that of a classic penalty.
Why are Italian SMEs exposed to this risk
Italian small and medium-sized enterprises often operate in competitive niches where differentiation occurs through comparative content, buying guides, and product reviews. These types of content are precisely those most at risk according to the new policy.
However, the problem isn't just about those who have knowingly adopted incorrect practices. Many SMEs rely on agencies or freelancers who produce content optimized with outdated logic. As a result, pages created even in good faith can fall into the categories flagged by Google.
According to McKinsey's analysis, the share of searches handled by AI systems in Search is expected to grow significantly by 2027-2028. Therefore, the risk of exclusion from AI responses will become progressively more relevant for overall organic visibility.
For SMEs that have invested in SEO strategy In recent years, this update requires a structured review of existing content. This is not a minor intervention.
The operational impact on content strategy
The new policy modifies the scope of what is considered quality content. In the past, a well-structured comparative page—with data, tables, and optimized anchor text—was considered a solid SEO asset. Today, if that same page is found to be built to artificially guide AI responses, it becomes a risky element.
Furthermore, Google has not provided an exhaustive list of prohibited patterns. This introduces a margin of interpretative uncertainty. Therefore, the most prudent strategy is to produce content that answers a real user question, with verifiable sources and without rhetorical contrivances.
In operational terms, the priority areas to monitor are as follows:
- Pages of the «best products» or «top 10» type built without transparent editorial criteria
- FAQ sections optimized with answers designed to appear in AI Overviews
- Content that systematically replicates queries associated with generative answers
- Texts with repetitive structures designed to saturate the context of LLMs
The team of SEO copywriting at SHM Studio is already updating its editorial frameworks to take these criteria into account.
What to do now: a step-by-step approach
First of all, an audit of existing content needs to be conducted. The goal is to identify pages that may fall into categories flagged by the new policy. Subsequently, an intervention priority must be established based on the traffic and strategic relevance of each page.
In addition to this, it is appropriate to review internal editorial briefs or those provided to external vendors. Content production must be geared towards real user utility, not the saturation of AI-associated queries. This principle is not new, but the new policy makes it more explicitly binding.
Finally, it is advisable to monitor Google Search Console more frequently in the coming weeks. Any manual or algorithmic penalty signals may emerge several weeks after the policy update date.
Who manages integrated campaigns? digital marketing It should also verify the consistency between organic and promoted content. The landing pages of Google Ads campaigns they are subject to the same qualitative assessments.
The still-open construction site: what Google hasn't clarified yet
The updated policy leaves some relevant questions open. In particular, it's unclear how Google distinguishes between content genuinely optimized for AI and content created to manipulate it. The dividing line is thin and, at the moment, left to algorithmic interpretation.
Similarly, it is not specified whether the penalties will only concern the presence in AI Overviews or also traditional organic rankings. This detail is crucial for those who have structured their visibility across both channels.
At SHM Studio, we are closely following the evolution of this policy. The operational guidance we provide to our clients is updated in real time, incorporating the analyses of our specialists. AI applied to marketing with Search Console data and official Google communications.
For those who wish to delve deeper into the topic of content quality from an AI perspective, SHM Studio's blog publishes regular updates in the section Resources and insights.
Outlook for 2027-2028
The May 2026 update is likely just the first in a series. The progressive integration of AI into search systems will make an internal Google regulatory framework increasingly necessary to manage attempts to manipulate generative models.
Therefore, SMEs that invest in editorially sound and verifiable content today will find themselves in a competitive advantage in the coming years. Conversely, those who continue to produce content optimized solely for traditional metrics risk losing visibility in both organic rankings and AI responses.
Strategies for SEO and of content marketing they must therefore evolve towards a more rigorous editorial model. This also applies to content distributed on channels such as LinkedIn, where consistency between organic presence and Google-indexed content is increasingly relevant.
Those who wish to start a review of their strategy can request an initial consultation through the page SHM Studio contacts. The team is available for a preliminary, no-obligation analysis.
Related articles
Discover other articles that explore similar topics in depth, selected to give you a more complete and stimulating view. Each piece of content is carefully chosen to enrich your experience.
