Google Reports 32% Increase in Malicious Prompt Injection Attempts on Public Web Content, Affecting AI Search Visibility
💻 Google published a security release on April 23, 2026, highlighting indirect prompt injections targeting AI systems on the web. Using Common Crawl data, Google found prompt injections for pranks, SEO manipulation, and more. A notable 32% increase in malicious detection was observed from November 2025 to February 2026. Indirect prompt injections are embedded in web content and can influence AI agents, posing a new challenge for technical SEOs to manage AI-input hygiene.
Share