Name: 如何让 AI 处于掌控之中
Uploaded: 2023-10-30T14:38:42Z
Duration: 12 min 10 s
Description: 近期令人振奋的商用、开源 AI 的涌现可能会在几年内带来可怕的超级智能 AI，顶尖研究人员和专家担心它会剥夺人类的权力或抹消人类。科学家迈克斯·泰格马克（Max Tegmark）带来了积极的愿景，分享了我们该如何掌控 AI，让它为我们工作，而不是我们为它工作。

近期令人振奋的商用、开源 AI 的涌现可能会在几年内带来可怕的超级智能 AI，顶尖研究人员和专家担心它会剥夺人类的权力或抹消人类。科学家迈克斯·泰格马克（Max Tegmark）带来了积极的愿景，分享了我们该如何掌控 AI，让它为我们工作，而不是我们为它工作。

如何让 AI 处于掌控之中

Artificial intelligence has advanced even quicker than many experts predicted, with systems like ChatGPT-4 signaling that artificial general intelligence could arrive within just a few years. For the first time, major industry leaders and government officials are openly warning about extinction-level risks from AI. But runaway AI is not inevitable. By designing provably safe systems—using techniques like formal verification—we can harness AI's power without sacrificing human control.