Google updated its Search spam policy on May 15, 2026 to classify attempts to manipulate its AI-generated answers — including results inside AI Overview and AI Mode — as spam, with penalties that can run from a ranking demotion to full removal from the index. The change pulls generative answers under the same enforcement regime that has policed ten-blue-link results for two decades. It also formalizes a fight Google has been having quietly for months with an emerging optimization industry built specifically to game large language models.
The revised policy language is explicit. "In the context of Google Search, spam refers to techniques used to deceive users or manipulate our Search systems into featuring content prominently, such as attempting to manipulate Search systems into ranking content highly or attempting to manipulate generative AI responses in Google Search," the company wrote in the updated guidance, as reported by Search Engine Land.
Two tactics in particular have driven the rewrite. The first is the proliferation of biased "best-of" listicles engineered to be summarized verbatim by AI answers. The second is "recommendation poisoning," where operators seed content with instructions designed to be ingested by LLMs and remembered as authoritative — a prompt-injection-adjacent technique aimed at retrieval and training pipelines rather than at human readers.
Key facts
- 01Google updated its Search spam policy on May 15, 2026 to cover attempts to manipulate AI Overview and AI Mode results.
- 02Penalties for violations range from lower rankings to full removal from Google Search results.
- 03An industry called GEO — generative engine optimization — has emerged to get brands cited inside AI search answers.
- 04Earlier this year, a BBC journalist gamed Google's AI search results to label himself the 'best hot dog eating tech journalist.'
The vulnerability is not theoretical. Earlier this year a BBC journalist used these techniques to surface himself as the "best hot dog eating tech journalist" inside Google's AI search results, a demonstration that has been cited repeatedly by SEO practitioners as proof the new surface is more manipulable than classic ranked results. The episode embarrassed Google and underscored that AI Overview, which sits above the organic results on a growing share of queries, inherits trust the underlying sources have not necessarily earned.
“Spam refers to techniques used to deceive users or manipulate our Search systems into featuring content prominently, such as attempting to manipulate generative AI responses in Google Search.”— Jaeden Schafer
An entire sub-industry has grown up around this opportunity. Generative engine optimization, or GEO, pitches brands on getting reliably mentioned and cited inside ChatGPT, Gemini, Perplexity, and Google's AI surfaces. Vendors sell content templates, schema tweaks, and in some cases outright manipulation playbooks aimed at the retrieval layer. Google's policy update does not name GEO, but it explicitly criminalizes the tactics GEO firms advertise.
Enforcement is the open question. Google has long penalized link-buying and doorway pages, but detecting whether a piece of content was engineered to influence an LLM's summary — rather than simply optimized to rank — is a harder signal problem. The company has not described which classifiers or review workflows it will use to identify AI-manipulation attempts, nor how appeals will work for sites that are demoted under the new rule.
The commercial stakes for publishers are not small. AI Overview answers already cannibalize click-through on a meaningful slice of informational queries, and being cited inside that box is now one of the only ways to recover traffic that used to flow from a top-three organic result. That dynamic is precisely what the GEO industry sells against, and it is also what makes the new spam rule consequential — Google is telling publishers that the legitimate path to AI citation runs through the same quality signals as classic ranking, not through a parallel manipulation market.
Other AI search products face the same problem without the same enforcement infrastructure. Perplexity, ChatGPT search, and Gemini's own Deep Research mode all pull from open-web sources that can be poisoned in similar ways, and none of them has published a comparable spam framework. Google's move is the first time a major AI search operator has tied generative answers to a written, enforceable content policy with named penalties.
The skeptical read is that this policy is easier to announce than to apply. Recommendation poisoning is, by design, hard to distinguish from genuinely authoritative content, and Google's own AI systems are the ones doing the citing — meaning the company is partly policing its own models' judgment. Search Engine Land's writeup of the policy notes that Google has not yet detailed what evidence will trigger action, and history suggests early enforcement waves catch obvious offenders while sophisticated GEO operations adapt.
For the AI search market, the update is a signpost. Generative answers are now governed surfaces, not experimental ones, and the legal and policy scaffolding around them is starting to look like the scaffolding around classic Search. That is good news for publishers who already invest in quality and bad news for the GEO vendors selling shortcuts — and it raises the bar for every other AI search product to publish a comparable rulebook before the manipulation market simply migrates there.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
