{
  "id": 7043,
  "url": "https://arxiv.org/abs/2603.18330v1",
  "title": "MemArchitect: A Policy Driven Memory Governance Layer",
  "summary": "Persistent Large Language Model (LLM) agents expose a critical governance gap in memory management. Standard Retrieval-Augmented Generation (RAG) frameworks treat memory as passive storage, lacking mechanisms to resolve contradictions, enforce privacy, or prevent outdated information (\"zombie memories\") from contaminating the context window. We introduce MemArchitect, a governance layer that decouples memory lifecycle management from model weights. MemArchitect enforces explicit, rule-based poli",
  "authors": "Lingavasan Suresh Kumar, Yang Ba, Rong Pan",
  "category": "research",
  "topics": "regulation,privacy-surveillance,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-03-18T22:37:05.000Z",
  "fetched_at": "2026-07-14T16:32:54.536Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/7043",
  "original_url": "https://arxiv.org/abs/2603.18330v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}