At Search Central Live Shanghai in May 2026, Google's Gary Illyes made something explicit that many in the GEO space had been quietly debating: buying or artificially manufacturing brand mentions to influence AI Overviews and AI Mode is something Google strongly warns against. Not a grey area. Not a 'proceed with caution.' A firm warning.
This matters because a small but growing cottage industry has emerged around exactly this tactic. Agencies and vendors offering to seed mentions of your brand across forums, low-quality editorial sites, and third-party roundups - specifically to feed AI systems the impression that your brand is widely discussed and trusted. Illyes' comments signal that Google is watching this space and does not consider it legitimate.
What 'Manipulating Mentions' Actually Looks Like
The tactics in question are not hard to spot once you know what to look for. Paid placements in 'best of' listicles without disclosure. Sponsored mentions disguised as organic editorial. Forum posts crafted to name-drop a brand in a context that looks like a genuine recommendation. Mass outreach to get a brand mentioned in AI-training-adjacent content. The goal in each case is the same: make the brand appear more cited and endorsed than it authentically is.
The parallel to early black-hat link building is obvious. Just as brands once bought links in bulk to manipulate PageRank, some are now buying mentions in bulk to manipulate AI citation patterns. Google has been here before, and its response has historically been the same: identify the signal being gamed, reduce its value, and penalise those exploiting it.
The specific callout of AI Overviews and AI Mode is significant. These are Google's own surfaces, and Google has both the motivation and the means to detect manipulation happening within its own ecosystem. If your mentions are coming from the same cluster of low-authority sites, appearing suddenly at volume, or lacking genuine engagement signals, that pattern is detectable.
The Problem with Shortcuts in AI Visibility
One reason these tactics have gained traction is that AI visibility is genuinely hard to measure and attribute. Brands see competitors appearing in AI Overviews and cannot always work out why. That uncertainty creates demand for quick fixes, and vendors are happy to supply them. The promise is simple: more mentions equals more citations. It is not that straightforward.
AI systems like those powering Google's AI Overviews are not simply counting mentions. They are assessing context, source credibility, corroboration across authoritative sources, and the consistency of information about a brand over time. A sudden spike in low-quality mentions does not replicate what genuinely authoritative coverage looks like. At best, it is ignored. At worst, it is a flag.
There is also a compounding risk. Brands that build AI visibility on manufactured mentions are building on an unstable base. When Google tightens its detection - and based on Illyes' comments, that appears to be a priority - brands relying on inauthentic signals face a harder reset than those who built visibility the slow way.
What This Means for GEO Strategy in Practice
The Illyes warning effectively confirms what sound GEO strategy has always pointed toward: AI visibility is earned through genuine authority, not engineered through volume. That means the work is fundamentally editorial and reputational rather than technical or transactional.
For brands actively working on AI Overviews visibility, this means prioritising placements and mentions that carry real editorial weight. A single mention in a well-regarded industry publication, a detailed product review on a high-trust comparison site, or a consistent presence in expert-authored content will do more than fifty mentions on sites that exist purely to publish sponsored text.
It also reinforces the value of owned content that AI systems can actually use. Detailed, accurate, well-structured content on your own site - supported by appropriate schema markup - gives AI systems something substantive to cite. Mentions elsewhere can point toward your brand, but the content on your own properties needs to be worth citing in the first place.
The Broader Implication for AI Search Across Platforms
Google's warning is specifically about its own surfaces, but the principle applies across the AI search field. Perplexity, ChatGPT, and Gemini all draw on signals of genuine authority and corroboration when deciding which brands and sources to surface. Manufactured mentions that might fool a volume-counting algorithm are unlikely to fool systems trained to assess source quality and information consistency.
The practical implication for UK brands is that AI visibility strategy needs to be approached with the same discipline as traditional media relations - building real relationships with credible publications, earning coverage through genuinely useful products and positioning, and ensuring that when AI systems look for information about your brand, what they find is consistent, authoritative, and accurate.
Illyes' comments also suggest that Google is actively developing its ability to distinguish authentic brand authority from engineered brand authority. That is not a threat to brands doing things properly. It is a threat to those who took shortcuts and expected them to hold.
A Clear Line Between Strategy and Manipulation
There is sometimes genuine confusion about where legitimate AI visibility work ends and manipulation begins. Optimising your content structure so AI systems can read it accurately is not manipulation. Ensuring your brand has a consistent, factually correct presence across authoritative sources is not manipulation. Pursuing genuine editorial coverage in relevant publications is not manipulation.
What crosses the line is paying for mentions that are presented as organic, seeding fake discussions to create the appearance of broader brand conversation, or coordinating placements specifically to game AI citation patterns rather than to genuinely inform audiences. The test is straightforward: would this content exist if AI search did not? If the honest answer is no, that is a signal you are working against the grain of what these systems are designed to reward.
Google has drawn a clear line. Brands that understand that line early are better positioned - both to avoid the risks of manipulation and to invest properly in the approaches that actually build sustainable AI visibility over time.