Name: 如何確保人工智慧在掌控中
Uploaded: 2023-10-30T14:38:42Z
Duration: 12 min 10 s
Description: 隨著目前令人振奮的商用人工智慧和開放程式碼人工智慧大爆發，在幾年內可能接踵而至的就是令人毛骨悚然的超智慧——研究人員和專家擔憂它可能會使人類失能或滅絕。科學家麥克斯‧泰格馬克會在這場演說中描述樂觀的願景，闡述了我們如何能夠控制人工智慧，確保它為我們工作，而不是相反。

隨著目前令人振奮的商用人工智慧和開放程式碼人工智慧大爆發，在幾年內可能接踵而至的就是令人毛骨悚然的超智慧——研究人員和專家擔憂它可能會使人類失能或滅絕。科學家麥克斯‧泰格馬克會在這場演說中描述樂觀的願景，闡述了我們如何能夠控制人工智慧，確保它為我們工作，而不是相反。

如何確保人工智慧在掌控中

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.