Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents

Abstract

Safety evaluations of memory-equipped LLM agents typically measure within-task safety: whether an agent completes a single scenario safely, often under adversarial conditions such as prompt injection or memory poisoning. In deployment, however, a single agent serves many independent tasks over a long horizon, and memory accumulated during earlier tasks can affect behavior on later, unrelated ones. Studying this regime requires evaluation along the temporal dimension across tasks: not whether an

Categories

Framework mappings

Suggested from the entry's categories.

@misc{altawaha2026remembering,
  title = {{Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents}},
  author = {Ahmad Al-Tawaha and Shangding Gu and Peizhi Niu and Ruoxi Jia and Ming Jin},
  year = {2026},
  month = may,
  eprint = {2605.17830},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.17830}
}