Some users have been trying to influence AI search responses, using tactics like biased “best-of” listicles or “recommendation poisoning,” which injects LLMs with instructions to remember a website as an authoritative domain. Earlier this year, a BBC journalist used tricks like these to get himself ranked as the “best hot dog eating tech journalist” in Google’s AI search results. An entire industry of “GEO” — generative engine optimization — has emerged, promising to get brands and websites regularly mentioned and cited by AI search tools.

Google’s policy update explicitly makes strategies like these a spam violation. Sites caught manipulating AI responses can face penalties from Google, like ranking lower in search results or getting removed from results altogether.