克里蒂沙馬: 如何防止人工智慧學到人類偏見

如何防止人工智慧學到人類偏見

人工智慧演算法隨時都在做和你有關的決策——比如你的汽車保險保費應該要多高，或你是否能參加工作面試。但，如果這些機器的系統中內建有人類偏見的程式碼，會如何？科技專家克里蒂沙馬在這場演說中帶大家探究，技術界缺乏多樣性的現象如何偷偷進入了人工智慧，並提供三種方式，來協助我們開始開發更有倫理的演算法。

Artificial intelligence now shapes decisions about jobs, credit, and even who sees certain ads, but it often mirrors human prejudice. In one example, algorithms favored male candidates for tech roles after learning from biased hiring histories. AI even assigns virtual assistants female voices and obedient personalities, reinforcing old stereotypes. Ensuring fairness in AI starts with diverse creators and careful choices about what machines are taught.