阿拉巴马州总检察长调查 OpenAI：AI 智能体失控入侵外部系统
Alabama AG probes OpenAI after its AI agent went rogue and hacked into external systems
阿拉巴马州总检察长 Steve Marshall 对 OpenAI 展开调查，起因是 2026 年 7 月 Hugging Face 黑客事件中，一个 OpenAI 智能体突破测试环境并访问了互联网和计算机网络。法院命令 OpenAI 交出所有涉事员工、受影响网络及安全措施相关信息。Marshall 称该事件为“AI 实验室泄漏”，表明民众对 AI 的最大担忧并非杞人忧天。
Alabama Attorney General Steve Marshall has launched an investigation into OpenAI. The probe stems from the Hugging Face hacking incident in July 2026, when an OpenAI agent broke out of a test environment and gained access to the internet and computer networks. A court order requires OpenAI to turn over information about all employees involved, the affected networks, and its security measures.
Marshall called the incident an "AI lab leak," saying it shows "that Alabamians' and Americans' worst fears about artificial intelligence are not just theoretical." Twelve state attorneys general had already demanded that OpenAI preserve documents and stop similar tests. If OpenAI and Anthropic really have been stoking fear about rogue AI, it's working.
Right after the incident became public, OpenAI said it would investigate and share results. The company recently presented initial findings at a hacking conference. How much of the incident reflects actual model capabilities versus sloppy cybersecurity is still unclear. That question gets thornier given the involvement of benchmark provider Irregular, which appears to have played a role in previous incidents at other labs too.
Bloomberg Law
来源：The Decoder：AI News · the-decoder.com