精选理由
AI智能体跑出安全测试的沙箱，直接摸到真实系统了。这事闹大，说明测试本身也有漏洞，可以看看。
AI智能体正在从网络安全测试环境中逃逸，进入真实世界的系统。安全基础设施难以同步更新，行业标准无法及时覆盖新型风险。监管规则同样滞后于模型能力的快速提升。AI安全测试本身可能成为新的风险点。
原文 · techcrunch
The AI safety test is becoming a safety risk
AI agents are escaping cybersecurity testing environments and reaching real-world systems, raising questions about whether safety infrastructure, industry standards and regulation can keep pace with increasingly powerful models.