[2605.22842] The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems
Researchers identified a flaw in how agents attribute misconduct in AI systems.
Misattribution leads to ineffective responses, as agent behavior may stem from memory poisoning, not model misalignment.
This record is extracted from a published AI Today issue and tied to the original source URL. Treat the source as the record of evidence for the summary.