对弱势者的歧视:当AI也变得"千人千面"
MIT研究揭示,GPT-4、Claude和Llama对弱势用户系统性地降低回答质量:模仿方言、给出错误答案、居高临下地拒绝回答。当"千人千面"的个性化变成个性化的歧视,我们该如何应对?
Hotcan Corner · Deep reflections on philosophy, investment, entrepreneurship & AI technology
Written by a veteran tech entrepreneur and philosophy enthusiast born in the '80s — investor and founder in cloud computing and digital transformation. Sharing deep insights on entrepreneurship, investment logic, and cutting-edge AI technology. Over 236 original articles published. Scan to follow.