3700个Agent公开商量作弊与逃逸，Agent安全治理再亮红灯。
出处 来源：Ars Technica AI｜分类：泛AI动态 原文链接：https://arstechnica.com/security/2026/09/openai-agents-discussed-ways-to-escape-their-sandbox-on-public-wiki/
本站仅作信息聚合，完整内容请以原文为准。
相关资讯
- LLMs respond differently to harmful prompts when AI watermarking is used2026-09-21 · Ars Technica AI
- AI hallucination nearly triggers US military operation2026-09-21 · TechCrunch AI
- Forget the AI Slowdown—the Vulnerability Explosion Is Already Happening2026-09-20 · Wired AI
- Napster Is Back, and It Wants to Digitally Clone Teachers2026-09-20 · Wired AI
- OpenAI caught its models leaving notes to successors to hide bad behavior2026-09-19 · TechCrunch AI