Core Insights - The article discusses the impact of AI on the job market, particularly focusing on how it affects entry-level positions and the employment landscape for young workers [6][7][8]. Group 1: AI's Impact on Employment - A study from Harvard University analyzed the effects of AI on the U.S. job market, revealing a significant decline in entry-level job opportunities since the introduction of AI technologies like ChatGPT [12][14][24]. - The research utilized a comprehensive dataset covering 285,000 companies and 62 million resumes, indicating a clear divergence in employment growth between junior and senior positions from 2015 to 2022 [17][18][19]. - The study found that companies adopting AI technologies saw a dramatic drop in entry-level hiring, with a 7.7% decrease in junior positions compared to non-AI companies after AI adoption [38][42]. Group 2: Mechanisms of Job Reduction - The decline in entry-level positions was primarily due to a halt in hiring rather than increased turnover rates, with AI-using companies hiring an average of 3.7 fewer junior employees per quarter [46][47][48]. - This trend was particularly pronounced in the wholesale and retail sectors, where AI-adopting companies reduced entry-level hiring by nearly 40% [55][56]. Group 3: Educational Background and Job Market Dynamics - The study categorized employees by their university tier, revealing a "U-shaped curve" where graduates from mid-tier universities (Tier 2 and Tier 3) faced the most significant job losses, while those from top-tier (Tier 1) and bottom-tier (Tier 5) universities were less affected [60][64][65]. - Companies preferred hiring top-tier graduates for their problem-solving abilities, while bottom-tier graduates were favored for their lower salary expectations, leaving mid-tier graduates vulnerable to AI-driven job reductions [63][64]. Group 4: Recommendations for Workers - Workers are advised to quickly transition from entry-level roles to more complex tasks that AI cannot easily replace, aiming for a job content where AI cannot substitute more than 50% of their responsibilities [69][74]. - Emphasis is placed on developing unique contextual knowledge and skills that AI cannot replicate, as well as leveraging personal interests to enhance competitiveness in the job market [75][80].

不少读者朋友给我留言，说能不能聊聊AI抢工作这件事。 这其实也反应了一种普遍的焦虑情绪，即大家隐约感觉AI会对工作造成冲击。 但它具体是怎么把一个办公室白领的饭碗给干掉的，很多人并没有真实的体感。 最近我在Twitter上看到了一篇非常精彩的论文，它全局、真切地研究了AI对工作的冲击。 我看了非常有感触，也分享给卫夕指北的读者。 论文来自哈佛大学，由两位经济学博士生Seyed M. Hosseini和Guy Lichtinger操刀。 而他们的导师是劳动经济学的重量级大咖拉里·卡茨（Larry Katz）。 因此，论文是严谨而有分量滴。 论文没有任何情绪渲染，就是用冰冷、庞大的真实数据，剖析了2023年以来美国就业市场的AI冲击具体是如何发生的—— 1 美国总共也就3.4亿人，所以，除掉老人和小孩，真正工作的人很大比例其实都包括在这个数据集里了。 结果一出来，一个清晰的剪刀差出现了—— 从2015年到2022年中，研究的目标公司数据集中，初级岗位（Junior-level）和高级岗位（Senior-level）的就业增长曲线，基本上是手拉手一起走的。 作为读研的时候也被写经济学论文折磨过的学术逃兵，在我看来， ...