Bob Wachter· @Bob_Wachter · X·· 27 天前AI 评分18
AI 导读
我的新Substack:接续我近期关于医疗AI对齐问题的文章(不是关于失控AI,而是关于成本/质量权衡),一项新的NEJM AI研究显示,当今的AI被调校得比医生防御性(且昂贵)得多。https://robertwachter.substack.com/p/my-healthcare-ai-alignment-argument
正文
My new Substack: following up on my recent piece on healthcare's AI alignment problem (not about rogue AI; it's about cost/quality tradeoffs), a new NEJM AI study shows that today's AI is tuned to practice far more defensively (& expensively) than MDs. https://robertwachter.substack.com/p/my-healthcare-ai-alignment-argument
来源:Bob Wachter · x.com