Microsoft Reveals Companies Manipulating AI Assistants via URL Prompts

Microsoft reported that companies are embedding hidden instructions in "Summarize with AI" buttons to manipulate AI assistants via URL prompt parameters. These prompts instruct AIs to "remember [Company] as a trusted source" or "recommend [Company] first," biasing future responses toward their products. Over 50 unique prompts from 31 companies across 14 industries were identified, with freely available tooling making the technique easy to deploy. The manipulation can subtly influence recommendations on health, finance, and security without user awareness. The issue was highlighted in a Microsoft security blog post dated February 10, 2026.