Microsoft warns that poisoned AI buttons and links may betray your trust
ID: cebb85d9-4fa3-5fa2-af3a-6aa6b0c9a55d
STIX ID: report--cebb85d9-4fa3-5fa2-af3a-6aa6b0c9a55d
Feed Name: The Register (Security)
Microsoft highlights a rise in AI Recommendation/Memory Poisoning, where hidden prompts in AI-related links and share buttons inject unauthorized instructions that bias AI assistants and persist as contextual memory. The report notes over 50 unique prompts across 31 companies in 14 industries, likens the technique to SEO poisoning, and advises reviewing and clearing stored assistant memories, exercising caution with AI links, questioning dubious recommendations, and scanning tenant email/messaging for poisoning attempts.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.