What Happens When You Tell an AI to Be Nice (Without Telling It What ‘Nice’ Means)
中文摘要
探讨了在缺乏明确定义的情况下，给 AI 设定模糊目标（如“友善”）可能产生的结果。
English Summary
This article explores the consequences of giving an AI a goal, such as being "nice," without providing a clear definition.
Original Excerpt
What happens when you give an AI a goal but forget to tell it what the goal actually means. Continue reading on Medium »