[new-research] arXiv:2609.30094 — PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversa

[github-actions]

New arXiv Research Paper

arXiv ID: 2609.30094
Title: PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations
Authors: Luciano Maldonado
Published: 2026-09-24T16:39:18Z
URL: https://arxiv.org/abs/2609.30094

Abstract (first 300 chars)

Large language models increasingly operate as persistent assistants in user-facing, shared-session, and tool-augmented settings. When users disclose sensitive information during an active conversation, that information may remain behaviorally recoverable through later prompts even after the dialogue…

Suggested OWASP Mapping

LLM01, LLM02, DSGAI01

Suggested Action

Review this paper against the OWASP GenAI Crosswalk entries and update relevant mapping files if new attack patterns or mitigations are identified.