Google just drew a hard line against the booming “generative engine optimization” industry. The company updated its spam policy to classify attempts to manipulate AI-generated search results, including AI Overviews and AI Mode, as spam violations, according to The Verge AI. Sites caught gaming the system can now face the same penalties Google reserves for traditional spam: ranking demotions or full removal from search results.
The policy change targets a specific set of tactics that have exploded over the past year. The Verge AI reports that bad actors have been using biased “best-of” listicles and “recommendation poisoning,” a technique that injects large language models with instructions to treat a website as an authoritative source. One BBC journalist showed how trivial this can be earlier this year, manipulating Google’s AI search to crown himself the “best hot dog eating tech journalist” on the internet.
What Google Actually Changed
The updated policy language is direct. Google now defines search spam to include “attempting to manipulate generative AI responses in Google Search,” putting it on equal footing with classic ranking manipulation. That’s a meaningful expansion. Until now, the spam rulebook was built for the blue-link era, focused on link schemes, cloaking, and keyword stuffing. AI Overviews lived in a gray zone.
Here’s what falls under the new rules:
- Crafting content designed to trick AI models into citing it as authoritative
- Injecting prompts or instructions meant to bias LLM outputs
- Publishing manipulated “best-of” lists engineered to be scraped by AI summarizers
- Any tactic aimed at gaming AI Overviews or AI Mode placement
Why This Matters
An entire cottage industry has sprung up around getting brands mentioned by AI search tools. They call it GEO, generative engine optimization, and agencies have been pitching it as the successor to SEO. Google’s update doesn’t ban GEO outright, but it draws a sharp line between legitimate optimization and manipulation. Agencies that promised guaranteed AI citations through prompt injection or recommendation poisoning just had their business model labeled spam.
What stands out here is the timing. Google has been racing to defend AI Overviews from accuracy complaints, including the infamous “put glue on pizza” episode. Letting bad actors poison those responses would compound the trust problem. Treating manipulation as spam gives Google’s enforcement teams a clear lever to pull.
What Practitioners Should Expect
If you’re running content for a brand, the practical implications are immediate:
- Audit your GEO vendor. If anyone you hired is using prompt injection or fake authority signals, you’re now exposed to ranking penalties.
- Focus on genuine authority. Google’s existing E-E-A-T framework still applies. Quality content that earns citations naturally is the only safe path.
- Watch for enforcement waves. Google typically rolls out spam policy changes alongside algorithmic updates. Expect manual actions and ranking adjustments to follow.
The broader shift is clear. AI search results are now a regulated surface, governed by the same spam principles as the rest of Google’s index. The free-for-all phase of GEO experimentation just ended.
Full details are available at The Verge AI.