Search
1 article for “reflective attribution collapse (RAC)”
-
TRACE: Tracking the Loss of Memory Provenance in LLM Agents
Abstract: AI agents that carry memory across sessions gain better personalization and decision-making, but this persistence opens a serious security gap. When an agent repeatedly condenses earlier interactions into compact “lessons,” the trail linking each lesson back to the interaction that produced it gradually fades. We term this effect Reflective Attribution Collapse (RAC): the progressive loss of provenance and forensic traceability that results from repeated memory reflection. Under RAC, a malicious …
Published in Journal of Advancements in Robotics · Vol. 13, Issue 2, 2026 · pp. 33–49 Read article