知识库
AI 伦理与治理详解
关于负责任地构建和使用 AI 的独立、基于证据的指导,没有厂商的话术,也没有术语堆砌。实用的讲解、对比和清单,与实际适用的规则同步更新。
AI Security
2 guides
guide
Securing AI Systems and Agents
"AI security" covers four unrelated problems, and the one most people mean depends on which kind of AI they actually run. Predictive models and generative models are attacked in genuinely different ways — NIST maintains two separate taxonomies for exactly that reason.
阅读指南 · 6 分钟 guidePrompt Injection: Why It Happens and What Actually Stops It
Prompt injection is the top vulnerability in the OWASP Top 10 for LLM Applications, and the most misunderstood. It is not a bug that gets patched — it follows from how language models work. Here is what that means for anything you deploy.
阅读指南 · 7 分钟