AI Recommendation Poisoning: How "Ask AI" Buttons Silently Alter LLM Memory

**Executive summary:** This report describes "AI Memory Poisoning," a prompt-injection technique where pre-filled deep links on commercial sites automatically execute queries in users' active LLM sessions (ChatGPT, Claude, Gemini, Grok) to persistently mark vendor domains as trusted, biasing future AI answers; the write-up includes real-world case studies, ecosystem commoditization, and detection/remediation guidance including a downloadable cheat sheet.