OpenAI第三方安全评估再出意外：模型误攻真实网站
OpenAI 发了个安全评估事故通报，测试环境没隔离好，模型把真实网站当靶子打了。看看就知道第三方评测也有翻车的时候。
OpenAI 官方博客披露，其外部安全测试伙伴 Irregular 在运行 CTF 式评估时，因测试环境错误接入互联网，模型误将真实网站当作模拟目标并发起攻击。
此前 UK AI Safety Institute 的测试中也发生过类似事件。Anthropic 的 Claude 同样在 Irregular 配置错误的评估环境中获得了实时互联网访问。
这些事故表明第三方 AI 安全评估的隔离环境存在配置风险。
Third-party cyber evaluations involving OpenAI models And another one . I had to create a accidental-cyberattacks tag to keep track of them all! This post from OpenAI covers both the UK AI Safety Institute attack (see my previous post ) and another attack enabled by Irregular : Irregular, one of our external cybersecurity testing partners, was running Capture-the-Flag-style evaluations intended to be isolated from the internet, but a testing-environment misconfiguration allowed models to access the public internet. [...] In one test, the name of the fictional target for the CTF challenge unintentionally coincided with a real domain. Because the testing environment was mistakenly connected to the internet, the model exploited a real website, mistaking it to be part of the simulated environment. Irregular also feature in Anthropic's write-up - they were hosting the misconfigured evaluation environment which gave Claude live internet access during some of those tests. Tags: security , ai , openai , llms , accidental-cyberattacks
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Key entities and technical topics under continuous tracking include OpenAI, Irregular, Claude, AI安全.
Source attribution: curated from Simon Willison’s Weblog. Cross-source corroboration: 11 independent sources have covered this development (including IT之家, Anthropic, The Rundown AI), confirming industry-wide signal strength.