📄 中文摘要
生成式 AI 代理和流氓交易员对雇主构成类似的内部威胁。企业在缺乏足够监督的情况下部署具有广泛影响力的代理 AI，这种情况为特定类型的风险提供了条件。文章分析了这种风险的表现形式，并提出了相应的应对策略，以帮助企业有效管理和降低潜在的内部威胁。通过加强监督和制定明确的使用政策，企业可以在利用代理 AI 的优势的同时，避免可能的滥用和损失。
识别和避免代理 AI 中的 ROT
出处: Spotting and Avoiding ROT in Your Agentic AI
发布: 2026年3月25日
📄 English Summary
Spotting and Avoiding ROT in Your Agentic AI
Generative AI agents and rogue traders present similar insider threats to their employers. Companies are likely to deploy agentic AI with extensive reach and insufficient oversight, creating conditions for a specific type of risk. The article analyzes the manifestations of this risk and proposes corresponding countermeasures to help organizations effectively manage and mitigate potential insider threats. By enhancing oversight and establishing clear usage policies, companies can leverage the advantages of agentic AI while avoiding potential misuse and losses.