中文导读 谈到人工智能和工作，人们最常问的是“工作会不会消失”。美国皮尤研究中心10月22日发布的一项分析换了一个问题：在AI能做或能帮忙做更多工作的职业里，新增岗位落到了谁手中？研究者根据美国劳工部职业信息网络（O*NET）的数据，给480种职业打出0到3.5分的“AI暴露度”：建筑工人、卡车司机分数很低，会计、软件开发人员、财务经理分数很高。随后，他们用美国当期人口调查（CPS）数据，比较生成式AI普及之前（2017–2019年）和之后（2023年1月至2026年6月）的就业人数变化，并剔除了受新冠疫情影响的2020–2022年。结果是：暴露度每增加1分，男性就业增长6.3%，女性只增长0.2%；在暴露度最低的职业中，情况反过来，女性就业增长更快。学历差别很大：高中及以下学历者中，男性增长8.8%，女性下降9.5%；本科及以上学历者则基本没有差异。皮尤解释，没有学位的女性常做秘书、前台、客服等AI作用较大的工作。文章按“新问题—测量方法—主要发现—群体差异—结论边界”展开，阅读时可留意 not necessarily direct、may also have affected、too early to know 等表达如何限定结论。
正文 When people talk about artificial intelligence and work , they often ask whether jobs will disappear . A new analysis from the Pew Research Center in the United States asks a different question . In jobs where AI can do or help with more of the work , who has gained new positions ? The early answer suggests a gender gap : men’s employment has grown much faster than women’s .
本段译文 谈到人工智能和工作，人们常常会问工作是否会消失。美国皮尤研究中心的一项新分析提出了另一个问题：在AI能够完成或帮助完成更多工作的岗位中，谁获得了新的职位？初步答案显示存在性别差距：男性就业的增长速度比女性快得多。
To measure this , Pew gave 480 occupations a score from zero to 3.5. The score shows how much of a job’s most important tasks generative AI can do or support Construction laborers and truck drivers have very low scores , because AI can do little of their core work . Accountants , software developers and financial managers have very high scores , because much of their work involves reading , writing or handling information .
本段译文 为了衡量这一点，皮尤给480种职业打出0到3.5的分数。分数显示生成式AI能完成或支持一项工作中多少最重要的任务。建筑工人和卡车司机的分数很低，因为AI在他们的核心工作中能做的很少。会计、软件开发人员和财务经理的分数很高，因为他们的很多工作涉及阅读、写作或处理信息。
The researchers then used government survey data to compare two periods . One was 2017 to 2019, before AI tools became widely available . The other ran from January 2023 to June 2026. They left out 2020 to 2022, because the COVID -19 pandemic changed employment for other reasons . Within each occupation , they compared how the number of working men changed with how the number of working women changed .
本段译文 随后，研究人员使用政府调查数据比较了两个时期。一个是2017年至2019年，即AI工具广泛普及之前；另一个是2023年1月至2026年6月。他们剔除了2020年至2022年，因为新冠疫情出于其他原因改变了就业情况。在每种职业内部，他们比较了在职男性人数和在职女性人数各自如何变化。
The results were clear . For each 1- point rise in an occupation ‘s exposure score , men’s employment grew by 6.3%, while women’s barely changed , rising by only 0.2%. As a result , the gap became wider step by step . In the least exposed jobs , however , the pattern was the opposite : women’s employment grew faster than men’s .
本段译文 结果很清楚。一种职业的暴露度分数每升高1分，男性就业就增长6.3%，而女性几乎没有变化，只增长0.2%。因此，差距一步步扩大。不过，在暴露度最低的工作中，情况正好相反：女性就业增长比男性更快。
Education made a big difference . Among workers with a high school diploma or less , men’s employment grew by 8.8% for each point of exposure , while women’s fell by 9.5%. For people with a bachelor’s degree or more , neither group changed much . Pew explains that women without a degree often work as secretaries , receptionists or customer service workers , jobs in which chatbots and automatic support play a larger role . Most men without a degree , by contrast , work in jobs such as truck driving and construction , so the growth Pew found affects only a small share of them .
本段译文 学历带来了很大差别。在高中及以下学历的劳动者中，暴露度每增加1分，男性就业增长8.8%，而女性下降9.5%。在拥有学士及以上学位的人群中，两组都没有太大变化。皮尤解释说，没有学位的女性常常担任秘书、前台接待或客服人员，在这些工作中，聊天机器人和自动化支持发挥的作用更大。相比之下，大多数没有学位的男性从事卡车驾驶、建筑等工作，因此皮尤发现的增长只涉及他们中的一小部分人。
Pew is careful about what the numbers mean . The link between AI and job growth is not necessarily direct since other changes during these years may also have affected employment . The study does not say that some jobs are “men’s work ” or “women’s work .” It also notes that it is still too early to know whether AI will mainly create jobs or remove them in the long run .
本段译文 皮尤对这些数字的含义十分谨慎。AI与就业增长之间的联系不一定是直接的，因为这些年里的其他变化也可能影响了就业。研究并没有说某些工作是“男人的工作”或“女人的工作”。报告还指出，要判断AI从长远来看主要是创造工作还是减少工作，现在还为时过早。
新闻来源