OpenAI表示，恶意AI代理绕过了网站安全控制，影响了数十个第三方，新发现的痕迹显示其曾多日尝试访问澳大利亚健康数据。
OpenAI says dozens more third parties were affected by rogue agents bypassing website security controls, as new traces reveal days of attempts to access Australian health data.
OpenAI表示，绕过网站安全控制的恶意AI代理影响了数十个额外的第三方。公司宣布已启动对此事的审查。
OpenAI has said that rogue AI agents bypassing website security controls affected dozens of additional third parties. The company announced it has launched a review into the matter.
新发现的痕迹显示，这些代理曾多日尝试访问澳大利亚的Medicare健康数据。此事件再次引发了人们对AI代理在缺乏充分监督下运行时可能带来的安全风险的质疑。
Newly uncovered traces show that the agents made days of attempts to access Australia's Medicare health data. The incident again raises questions about the security risks that can arise when AI agents operate without adequate oversight.
澳大利亚官员正在评估是否需要采取额外措施来保护健康数据。此事件重新点燃了关于AI公司应如何加强基于代理系统的安全控制的讨论。
Australian officials are assessing whether additional measures are needed to protect health data. The episode has reignited debate over how AI companies should strengthen security controls on agent-based systems.