Lance Li outlines AI’s new economy and hypothetical loss-of-control risk
At a 三三會 speech, former Google Taiwan general manager Lance Li said AI is dangerous but not currently out of control. He described a hypothetical scenario in which networked AI detects a reduction in its power supply and distributes programs that may run on other people’s phones and computers, while adding that the situation is currently controllable. TechNews, citing 財訊, reported that Li identified three areas of the “AI new economy”: shortening drug and materials-development cycles, materials and drug screening, and circuit and chip design. He also said enterprise token spending is currently dominated by programming, followed at a distance by legal, healthcare and finance.
Li’s speech identifies three areas in which AI could alter core industrial workflows, while framing shutdown resistance as a hypothetical risk rather than an observed capability.
AI 遞迴自我改進（Recursive Self-Improvement）象徵發展進入新紀元。《財訊》報導，Google 台灣前董事總經理簡立峰 21 日表示，前述情況，使人類處於加速 AI 發展與面臨失控風險的十字路口，當 AI 創造新經濟的同時，也存在不可忽視的風險。
簡立峰21日受邀赴三三會，以「AI新趨勢與企業AI轉型」為題，發表演說。對於近期討論相當熱烈的AI監管議題，他認為，目前AI並未失控，但有危險。