Brookings study finds no measurable pay or hours effect from generative AI in Denmark
Two Brookings Institution economists analyzed about 25,000 Danish workers across 11 AI-exposed occupations, linking survey responses from late 2023 and late 2024 to monthly administrative records through December 2024. They report no measurable change in earnings, hours or workplace employment for chatbot users or adopting workplaces, with statistical precision sufficient to rule out differences above 2%. Work content was substantially reorganized, new tasks such as AI integration and output review emerged, and adopters recorded about 4% more full-time-equivalent hours in their latest occupation than similar non-adopters, which the researchers interpret as a greater likelihood of having changed occupation after chatbots appeared. Movers into more AI-related, higher-paying roles saw income growth 12 percentage points ahead of other Danish workers.
The authors suggest the adjustment is occurring inside workplaces through task reorganization rather than in wages or hiring, echoing the productivity J-curve literature in which organizational investment precedes measurable economic results.
导语ChatGPT问世以来，生成式AI正加速渗入白领的日常工作。但它究竟是在替代劳动者，还是帮其提升能力、转向更高薪的岗位？布鲁金斯学会（Brookings Institution）的两位经济学家追踪2.5万名丹麦劳动者后给出答案：生成式AI尚未对工资、工时与企业用工产生可测量的影响，但工作内容却被实质性重组......
本文来自微信公众号： 数字经济发展评论 ，作者：数字经济发展评论
导语
ChatGPT问世以来，生成式AI正加速渗入白领的日常工作。但它究竟是在替代劳动者，还是帮其提升能力、转向更高薪的岗位？布鲁金斯学会（Brookings Institution）的两位经济学家追踪2.5万名丹麦劳动者后给出答案：生成式AI尚未对工资、工时与企业用工产生可测量的影响，但工作内容却被实质性重组，部分使用者已经转入了薪酬更高的职业。