| 96 | 0 | 97 |
| 下载次数 | 被引频次 | 阅读次数 |
近年来，生成式人工智能不断取得突破，颠覆性的技术进步在推动经济社会发展的同时，也带来了一些新的挑战，对就业市场的冲击就是其中之一。科技进步与就业之间潜藏着复杂的关系，新兴技术既能推动产品（服务）的原始创新，从而扩大就业规模，促进充分就业；也能推动生产（服务）技术创新，产生劳动力替代效应，导致劳动力需求缩减。新一代信息通信技术对生产（服务）技术创新的推动作用更为显著，其中人工智能技术产生的劳动力替代效应尤为显著。因此，不能简单地根据历史经验认定新技术在消灭旧工作岗位的同时，会创造出等量的乃至更多的新工作岗位。目前，我国三次产业均已呈现无就业经济增长态势。当前生成式人工智能还处于泛行业应用阶段，其劳动力替代效应尚不显著。但生成式人工智能正循垂直大模型、AI智能体、具身智能三种应用进路在各个行业和细分领域普及，相关产品呈现智能化水平持续提升、成本不断下降趋势，其劳动力替代效应将不断增强。在可预见的未来，随着千行百业的智能化转型升级，第一产业、第二产业的就业规模将会继续缩小，而第三产业的众多职业类别也将借助生成式人工智能实现半自动化或自动化。新兴技术与就业之间潜藏着复杂的关系，不能简单地根据历史经验认定新技术在消灭旧工作岗位的同时，会创造出等量的甚至更多的新工作岗位。应对冲击的关键性策略是将产品（服务）创新放在首要位置，以拓展就业空间，促进充分就业。
Abstract：In recent years, generative artificial intelligence has continuously made breakthroughs. While these disruptive technological advances drive economic and social development, they also bring new challenges — including impacts on the job market. The relationship between technological progress and employment is complex. On the one hand, emerging technologies can foster product(or service) innovation, thereby expanding employment and promoting full employment. On the other hand, they can also drive production(or service) process innovation, generating labor substitution effects and reducing labor demand. The new generation of information and communications technology plays a particularly significant role in driving production(or service) process innovation, and among these technologies, artificial intelligence exhibits the most pronounced labor substitution effect. Therefore, we cannot simply assume from historical experience that new technologies will create an equal or even greater number of new jobs while eliminating old ones. Currently, all three industries in China have exhibited a trend of jobless growth. Generative AI is currently still at the stage of cross-industry application, and its labor substitution effect is not yet pronounced. However, generative AI is spreading across industries and sub-sectors through three application trajectories: vertical-specific large models, AI agents, and embodied intelligence. With steadily rising intelligence and falling costs, its labor substitution effect is set to become increasingly pronounced. In the foreseeable future, as a wide range of industries undergo intelligent transformation and upgrading, employment in the primary and secondary industries will continue to shrink, while many job categories in the tertiary industry will also achieve semi-automation or full automation through generative AI. However, the relationship between emerging technologies and employment is complex. We cannot simply assume from historical experience that new technologies, while destroying old jobs, will automatically create an equal or even greater number of them. The key strategic response to this challenge is to prioritize product(or service) innovation, thereby expanding the space for employment and promoting full employment.
(1)曾晨语，魏下海，余玲铮.生成式人工智能如何重塑劳动力市场：研究综述与未来展望[J].产业经济评论，2025(3):82-102.
(2)易宪容，陈颖颖.生成式人工智能对劳动力市场的颠覆性影响及其应对[J].江海学刊，2024(6):91-99.
(3)无就业经济增长：又称零就业经济增长，是描述经济增长未伴随就业增长现象的经济学术语。该概念于2020年经全国科学技术名词审定委员会审定发布，定义为经济总量扩大但就业岗位未同步增加的状态。
(1)杨胜利，王伟荣.劳动就业与收入分配[M].武汉：武汉大学出版社，2023:36.
(2)熊彼特.经济发展理论：对于利润、资本、信贷、利息和经济周期的考察[M].何畏，易家详，等译.北京：商务印书馆，2020:76.
(1)陶凤，吕银铃.外贸回稳向好带动就业人数达1.8亿[N].北京商报，2020-10-15(2).
(2)国家统计局，国务院第四次全国经济普查领导小组办公室.第四次全国经济普查公报（第四号）——第三产业基本情况之一[EB/OL].(2019-11-20)[2025-07-25]. https：//www.stats.gov.cn/sj/zxfb/202302/t20230203_1900527.html.
(3)国家统计局，国务院第五次全国经济普查领导小组办公室.第五次全国经济普查公报（第四号）——第三产业基本情况之一[EB/OL].(2024-12-26)[2025-07-25]. https：//www.stats.gov.cn/xxgk/sjfb/zxfb2020/202412/t20241226_1957894.html.
(1)SCHAEFFER R, MIRANDA B, KOYEJO S. Are emergent abilities of large language models a mirage?[J]. Advances in neural information processing systems,2023(36):55565-55581.
(1)查睿，刘惠宇. DeepSeek-R1幻觉率高达14.3%，聪明的AI为何总爱胡说八道？[EB/OL].(2025-02-28)[2025-07-26]. https：//export.shobserver.com/baijiahao/html/867068.html.
(2)文乐乐. AI幻觉可能越来越严重[N].中国科学报，2025-05-13(2).
(3)OpenAI新推理AI模型幻觉问题更严重[EB/OL].(2025-04-22)[2025-07-26]. https：//www.thepaper.cn/newsDetail_forward_30695382.
(1)LI P F, YANG J Y, ISLAM M A, et,al. Making AI less‘thirsty’:uncovering and addressing the secret water footprint of AI models[J]. Communication of the ACM,2025(7):54-61.
(2)文巧.大模型太“烧钱”！OpenAI据称今年将亏损50亿美元，算力成本或高达70亿美元，占总运营成本八成以上[EB/OL].(2024-07-26)[2025-07-26]. https：//www.nbd.com.cn/articles/2024-07-26/3481106.html.
(3)岳楚鹏，兰素英.“DeepSeek或彻底改变游戏规则”！中国大模型“搅动”硅谷，巨头进入恐慌模式，外媒刷屏，大佬发声：中国AI已追上美国[EB/OL].(2025-01-26)[2025-07-26]. http：//www.nbd.com.cn/articles/2025-01-25/3737591.html.
(4)陈奕桐. OpenAI首席执行官：DeepSeek-R1是一款令人印象深刻的模型，有这样的新竞争对手令人振奋[EB/OL].(2025-01-28)[2025-07-26]. https：//world.huanqiu.com/article/4LFpUXq5XCw.
(1)冯烁.DeepSeek什么来头，何以震动全球AI圈[EB/OL].(2025-01-28)[2025-07-26]. http：//www.ce.cn/xwzx/gnsz/gdxw/202501/28/t20250128_39280352.shtml.
(2)Stanford Institute for Human-Centered Artificial Intelligence. The 2025 AI index report[R/OL].[2025-07-26]. https：//hai.stanford.edu/ai-index/2025-ai-index-report.
(3)中国互联网络信息中心.生成式人工智能应用发展报告（2025）[R/OL].(2025-10-23)[2025-11-26]. https：//www.cnnic.net.cn/n4/2025/1021/c88-11391.html.
(1)华为云盘古大模型5.5概览[EB/OL].[2025-07-25]. https：//www.huaweicloud.com/product/pangu.html.
(1)周鸿祎：未来一个员工领导100个智能体将成常态[EB/OL].(2025-06-11)[2025-07-26]. https：//tech.ifeng.com/c/8k6bXcmnaoU.
(2)胡含嫣.百度文心智能体平台开发者突破80万，智能体数量自今年6月起增长310%[EB/OL].(2024-11-12)[2025-07-26].https：//m.thepaper.cn/baijiahao_29321571.
(1)庄祥贵，张宇.应用潜力升级！“绵阳造”人形机器人接入华为盘古大模型[EB/OL].(2024-06-22)[2025-07-26]. https：//www.scjjrb.com/2024/06/22/99404136.html.
(2)孙凝晖.人工智能与智能计算的发展[EB/OL].(2024-04-30)[2025-07-25]. http：//www.npc.gov.cn/npc/c2/c30834/202404/t20240430_436915.html.
(1)苏雁，李健.为农业生产插上智慧翅膀——江苏“无人化”农场建设见闻[N].光明日报，2025-05-11(3).
(2)陈辉.给农田装上“AI大脑”！中国移动在豫发布“万象耕耘”农业大模型[EB/OL].(2025-04-18)[2025-07-27]. https：//app-api.henandaily.cn/mobile/view/news/434291594749804544527107.
(3)郭旺，杨雨森，吴华瑞，等.农业大模型：关键技术，应用分析与发展方向[J].智慧农业（中英文），2024(2):1-13.
(1)国务院.国务院关于印发《中国制造2025》的通知：国发[2015]28号[A/OL].(2015-05-19)[2025-07-25]. https：//www.gov.cn/zhengce/content/2015-05/19/content_9784.htm.
(2)BILL M,MÜILLER C,KRAUS W,et al.World robotics 2024[R/OL].(2024-09-24)[2025-07-27]. https：//ifr.org/img/worldrobotics/Press_Conference_2024.pdf.
(3)国家矿山安监局，应急管理部，国家发展改革委.关于印发《关于深入推进矿山智能化建设促进矿山安全发展的指导意见》的通知：矿安[2024]42号[A/OL].(2025-11-13)[2025-11-26]. http：//cq.chinamine-safety.gov.cn/mjgov/html/ksaqxxhznhjs2/20251113/83427.html.
(4)国家电网发布国内首个千亿级多模态电力行业大模型[EB/OL].(2024-12-21)[2025-07-25]. http：//www.xinhuanet.com/energy/20241219/caa68c8595384ca0b4a49ae5bdc34a22/c.html.
(5)戴小河，王希.发电行业大模型“擎源”亮相[EB/OL].(2025-07-01)[2025-07-25]. https：//www.news.cn/fortune/20250701/c265ea6f0f8a47f7ac41390ddc7b8f47/c.html.
(6)陈杰. AI大模型填补供热行业智能化空白[EB/OL].(2025-03-27)[2025-07-25]. https：//www.stdaily.com/web/gdxw/2025-03/27/content_315661.html.
(7)郎佳慧，吕少威.机器狗、大模型……AI如何让燃气产业“燃”起来？[EB/OL].(2025-05-20)[2025-07-26]. https：//vs.cns.com.cn/views/details/details.html?id=585535.
(1)业内首个！百度智能云发布水业大模型白皮书[EB/OL].(2024-03-28)[2025-07-26]. https：//cloud. baidu. com/news/news_cf460b0c-0c58-4c88-8a03-a22a5d89bbae.
(2)陈雯.从答题到推理，人工智能将重塑建筑行业生产力[N].建筑时报，2025-02-24(2).
(3)孔祥利，莫正晖.生成式人工智能对我国服务业就业影响研究[J].陕西师范大学学报（哲学社会科学版），2025(1):102-114.
(4)北京大学国家发展研究院，智联招聘.AI大模型对我国劳动力市场潜在影响研究报告（2024）[R/OL].(2024-10-11)[2025-07-27]. https：//baijiahao.baidu.com/s?id=1810229677929813378&wfr=spider&for=pc.
(1)泰盛，解维奇.奥特曼：有些工作岗位已彻底淘汰，就像AI客服电话反而更方便[EB/OL].(2025-07-23)[2025-11-26]. https：//m.thepaper.cn/newsDetail_forward_31240441.
(2)胡谦.智能分拣、无人配送车“上岗”，助力成都快递业迎战“史上最长双11”[EB/OL].(2024-10-31)[2025-07-25]. https：//baijiahao.baidu.com/s?id=1814444697438710295&wfr=spider&for=pc.
(3)国家统计局，国务院第四次全国经济普查领导小组办公室.第四次全国经济普查公报（第四号）——第三产业基本情况之一[EB/OL].(2019-11-20)[2025-07-25]. https：//www.stats.gov.cn/sj/zxfb/202302/t20230203_1900527.html.
(4)国家统计局，国务院第五次全国经济普查领导小组办公室.第五次全国经济普查公报（第四号）——第三产业基本情况之一[EB/OL].(2024-12-26)[2025-07-25]. https：//www.stats.gov.cn/xxgk/sjfb/zxfb2020/202412/t20241226_1957894.html.
(1)黄旭，刘红英，闫雪凌.生成式人工智能的就业效应与应对策略[J].当代经济管理，2025(4):73-87.
(2)布什，霍尔特.科学：无尽的前沿[M].崔传刚，译.北京：中信出版集团，2021:54.
(3)司托克斯.基础科学与技术创新：巴斯德象限[M].周春彦，谷春立，译.北京：科学出版社，1999:63.
(1)熊彼特.资本主义、社会主义与民主[M].吴良健，译.北京：商务印书馆，1999:155.
基本信息:
DOI：10.16524/j.45-1002.2026.02.003
中图分类号:F249.2;TP18
引用信息:
[1]罗俊.生成式人工智能的劳动力替代效应及对充分就业的影响[J].学术论坛,2026,49(02):38-56.DOI:10.16524/j.45-1002.2026.02.003.
2026-03-26
2026-03-26