Utility Under Attack: Agent Memory Poisoning and the Limits of Content Screening and Provenance Ranking

In 30 Sekunden

Was passiert ist

Researchers demonstrated that poisoning just 1.2% of an AI agent's memory with false statements drops accuracy from 85% to 30%, and content screening fails to catch it.

Warum es zählt

Matters for engineers building AI systems with persistent memory or retrieval-augmented generation that store and reuse information across sessions.

Achtung

Current defenses like content screening and provenance weighting either fail entirely or create unusable tradeoffs between blocking poison and allowing legitimate untrusted sources.