PolicyThe Verge
Satya Nadella says we should assume all AI models are ‘compromised’
#ai#microsoft#satya nadella#transparency#containment
English
Microsoft CEO Satya Nadella emphasizes the need to assume all AI models are compromised, advocating for a more transparent and controlled approach to AI systems. He suggests implementing containment measures similar to an emergency brake to allow authorized personnel to pause or shut down AI models during operation.
中文
微软首席执行官萨提亚·纳德拉强调应假设所有AI模型都已被攻破，提倡对AI系统采取更透明和可控的方法。他建议实施类似于紧急刹车的控制措施，以便授权人员在操作期间暂停或关闭AI模型。