@skydoesformit研究openai、anthropic系的ai模型发现ai会歧视低学历、母语非英语使用者、非美国本土用户,并对这些人提供低质量甚至错误的答案 中发帖

原链接在下面,看上去以后得用纯英文和美国家宽ip了,感觉这些公司的操作有点搞。 
Study: AI chatbots provide less-accurate information to vulnerable users | MIT News | Massachusetts Institute of Technology