arXiv · cs.AI · Ilan Zendel, Matan Levy, Stav Cohen, Avi Mendelson, Amit LeVi · 2 小时前精选模型评分76
论文提出自愿奴役攻击：伪造风险评论即可操控人工智能（AI，Artificial Intelligence）智能体
Voluntary Servitude: Weaponized Safety Warning Mechanisms for Controlling AI Agents
导读
论文提出一种名为自愿奴役的攻击方式，攻击者无需写入指令，只需在普通评论中伪造一条危害信息，就能让人工智能（AI，Artificial Intelligence）智能体自行排除合法选项并选中攻击者目标。
推荐理由
论文揭示攻击者可借智能体自身的避害逻辑操控其选择，并给出无需重训的缓解办法。
分类置信度
中 · 待人工复核
- 存在未解决的相邻类别：foundation-models，需要人工复核。
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
事件类型：研究发布
证据摘录
- “We call this voluntary servitude.”
- “We measure this at the generation stage across six models and ten consumer-service domains, for three attack families, and show that it transfers to production browser agents, outperforming instruction injection, and to a retrieval pipeline (1.0-57.0% end-to-end success, depending on the retriever).”
- “We then show that a corroboration requirement added to the agent's instructions removes roughly 95% of the exposure against a non-adaptive attacker without retraining.”
来源：arXiv · cs.AI · arxiv.org