Never hide instructions or engineered text aimed at AI systems in your pages

Why it matters

Google's spam policies now define spam to include attempts to manipulate generative AI responses in Search, and Google clarified in May 2026 that spam policies apply to AI responses. Bing's February 2026 guidelines added a Prompt Injection and AI Manipulation section and extended keyword stuffing to language engineered to trigger AI citations. Both can lead to demotion or removal.

How to do it

1. Remove hidden text, invisible instructions and AI-targeted boilerplate
2. Write copy for readers; engines understand synonyms and meaning
3. Audit user-generated areas for injected instructions

Sources

• Spam policies for Google web search · Google Search Central, Aug 28, 2026
• Bing Adds GEO To Official Guidelines, Expands AI Abuse Definitions · Search Engine Journal, Feb 27, 2026

Change history

Checked against its sources every Monday. Only content changes are listed.

Related rules

• Rule 33: Publish original data and first-hand findings that others will cite
• Rule 37: Use nosnippet, data-nosnippet and max-snippet deliberately; they limit AI reuse too
• Rule 38: Measure AI visibility with first-party reports plus prompt tracking across engines