Google classifies attempts to manipulate AI answers in Search as spam

Google has updated its search spam rules and explicitly included attempts to manipulate generative AI answers. The new wording applies not only to regular results but also to AI Overview and AI Mode in Google Search.

The new wording

Google now treats as spam not only schemes that push pages up in classic results, but also attempts to influence what exactly the built-in AI in Search will say. The policy states plainly that spam includes techniques that deceive users or manipulate search engines to make content more visible, including influence on generative answers. This is an important clarification, because AI Overview and AI Mode are gradually becoming a separate entry point to the internet rather than a mere addition to a list of links.

"Spam is techniques that deceive users or manipulate search engines."

Previously, Google's anti-spam rules were mostly read through the lens of SEO and page ranking. Now the company states outright that the same logic applies to AI answers inside Search. This is not a cosmetic rewording: Google is showing that it regards generative blocks as part of the main results, and so wants to protect them from artificial pressure just as it protects regular results. If someone tries to promote not a page but the model's output itself, that also counts as a violation.

How answers are being influenced

The reason for the update is clear: a separate layer of gray-area practices has already appeared around AI search. Publishers, SEO specialists and marketers are trying not just to win a top spot but to embed themselves directly in the model's answer. They use biased roundups of the "best" services, content with recommendations baked in beforehand, and so-called recommendation poisoning, where pages are deliberately filled with phrasing that an LLM may pick up as seemingly neutral advice. On the surface it looks like an ordinary article, but in practice it works as pressure on the AI answer.

• roundups of the "best" products where the winner is decided in advance
• texts that disguise advertising as an independent recommendation
• pages written not for the reader but to be cited by a model
• repeated phrasing and comparisons that nudge the AI toward a desired conclusion

The problem is that the user may not open a single link at all. If AI Overview shows a ready-made answer right away, manipulation at the level of source material turns into manipulation of the user's actual decision. In such a scheme it matters less who ranks first in the familiar results. What matters more is whom the model names as the best, most recommended or most suitable option, and that is where a new fight is now beginning.

Why it matters

For Google, the issue is not only search quality but also trust in the new format. The company promotes AI Overview and AI Mode as a fast way to get a condensed answer without browsing many links. But the format becomes vulnerable if more aggressive players learn to systematically feed the model the conclusions they want. Users would then see not the most useful or accurate result, but the one whose authors have best mastered manipulation for generative results. For a search engine this is a direct risk: lost trust in AI features would hurt the whole product.

The update also looks like a signal to the SEO market. Content optimization as such is not going away: sites will keep working on the structure, clarity and authority of their material. But Google draws the line where deception of the user or the system for artificial advantage begins. In other words, it is no longer about making material easier for a search engine to understand, but about trying to force a model to repeat a recommendation that benefits someone. After this clarification, gray-area experiments around AI Search carry a much more explicit risk of sanctions.

What it means

Google has formalized a new front in the fight for search traffic: spam now covers not only attempts to lift a page to the top but also attempts to force through the AI answer itself. For brands and SEO teams, this signals that the fight for visibility in AI search will run under the same anti-spam rules as regular results. For users, it is a chance to get less skewed AI summaries, provided Google can apply its own rules consistently.