Nadella：应假定所有 AI 模型都已被攻破，须设紧急刹车
微软 CEO 主张默认每个 AI 模型都已被攻破、必须能随时被叫停并留下可审计证据——这类安全诉求正进入大公司话术。
微软 CEO Satya Nadella 在 X 发长文称，不能再把 AI 当作只能接受或拒绝其建议的"嵌套黑盒"。他主张从一开始就假定模型已被攻破并加以隔离，装一道"紧急刹车"——授权者随时能暂停或关闭任务中的模型，同时推动事故及时披露、独立审计和可验证数据。
正文摘录
- title: Satya Nadella says we should assume all AI models are ‘compromised’ - sourcecompany: The Verge - bodymarkdown: - AI AI Posts from this topic will be added to your daily email digest and your homepage feed. Follow [See All AI](https://www.theverge.com/ai-artificial-intelligence) - News News Posts from this topic will be added to your daily email digest and your homepage feed. Follow [See All News](https://www.theverge.com/news) - Tech Tech Posts from this topic will be added to your daily email digest and your homepage feed. Follow [See All Tech](https://www.theverge.com/tech) Satya Nadella 说，我们应当假定所有 AI 模型都已被 "攻陷" Microsoft 的 CEO 呼吁给 AI 装上一个 "紧急刹车"。 Microsoft 的 CEO 呼吁给 AI 装上一个 "紧急刹车"。