2025年AI职场报告：生成式AI如何重塑职业的DNA

The future of work and the role of generative artificialintelligence (GenAI) is not a question of whetherGenAI will change jobs, but what kinds of jobs willbe most and least changed, why, and how. Authors: Annina Hering, Arcenis Rojas www.hiringlab.org Indeed’s AI at Work is a series of research reports from Hiring Lab,Indeed’s economic research arm. This research is designed toprovide data and new insights into the impact of AI on the labormarket to help employers and job seekers better understand andprepare for the changing workforce. Annina Hering, Ph.D. Annina Hering is a senior economist at theIndeed Hiring Lab, where she bridges economics,data science, and data engineering to uncoverlabor market insights. She was the principalresearcher forHow GenAI is Rewiring the DNA ofJobs. Annina earned her Ph.D. in social sciencefrom the University of Cologne. Before joiningIndeed, she was a postdoctoral researcher at theMax Planck Institute for the Study of Societies(MPIfG) and a member of the International MaxPlanck Research School on the Social and PoliticalConstitution of the Economy (IMPRS-SPCE).She also holds a master’s degree in sociologyand empirical research from the University ofCologne and a bachelor’s degree in politics andsociety from the University of Bonn. Indeed Hiring Lab creates innovative data insights on the globallabor market that inspire new conversations about the state ofwork. As the economic research arm ofIndeed, the world’s numberone job site, Hiring Lab is driven by a team of leading economistsand data scientists who provide thought leadership on global labormarket conditions, including hiring trends, salary information,popular skills, and employer benefits. Hiring Lab analyzes millionsof data points across time collected from Indeed’s job postings,resumes, and job seeker behaviors to reveal emerging trends inthe United States and across the world. The unique insights generated by Hiring Lab inform talentmanagement, employment, and labor policy decisions forbusinesses, researchers, academics, and job seekers alike. HiringLab partners with a range of policy-making organizations andNGOs including the International Monetary Fund, the EuropeanCentral Bank, and the Bank of Canada to produce timely, incisiveresearch. Hiring Lab data is also regularly cited in prominentbusiness publications such as The Wall Street Journal, CNN,Reuters, The Globe and Mail, Der Spiegel, and The Financial Times.Hiring Lab economists regularly speak about labor market trendsat leading industry, policy, and academic conferences. Arcenis Rojas Arcenis is a data scientist at Hiring Lab. He hasa Bachelor’s in Applied Economics from IthacaCollege and a Master’s in Economics from the CityUniversity of New York — the city college wherehe found a love for working with economic data.Arcenis previously worked at the U.S. Bureau ofLabor Statistics where he gained proficiency inprogramming in R and coordinated activities forthe R Users’ Group. Later he went on to work as adata science consultant for U.S. federal agencies. For more information, please visit www.hiringlab.org September 2025 KEY POINTS More than a quarter (26%) of jobs posted on Indeed in thepast year could be “highly” transformed by GenAI. But themajority (54%) are likely to be “moderately” transformed,and their evolution will depend on how quickly businessesadopt GenAI, and how well workers adapt and reskill. Almost half (46%) of skills in a typical US job posting arepoised for “hybrid transformation” by GenAI. Humanoversight will remain critical when applying these skills,but GenAI can already perform a significant portion ofroutine work. Some occupations, including software development,are more highly exposed. Roles requiring more physicalpresence and human interaction, including nursing, arelikely to be less impacted, with GenAI primarily changingadministrative tasks but not necessarily the core parts ofthese jobs. The Indeed GenAI Skill Transformation Index measureshow much GenAI could change the way different skillsor jobs are done. Instead of measuring if GenAI iscapable of fully replacing a human worker, it examineshow skills may be applied going forward and howhuman involvement with those skills and tasks maychange. GenAI models evaluated both the cognitiveand physical demands required across almost 2,900work skills, and GenAI’s capacity to perform them.Based on this evaluation, skills were grouped intofour distinct categories based on their potential tobe transformed by GenAI: Minimal transformation,assisted transformation, hybrid transformation, andfull transformation. Earlier versions of this analysis found that zero humanwork skills were “very likely” to be fully replaced by GenAI.Today, 19 skills (0.7% of all skills analyzed) were assessed tobe “very likely” to be fully replaced by GenAI, still small inabsolute terms, but a significant signal of progress. Critically, these are only measures ofpotentialtransformation. Any realize