
一场无声的危机,正在动摇现代社会的根基。
支撑全球经济运转的工业工人队伍,正在面临瓦解风险。负责保障电网稳定运行、工厂持续运转、公用事业可靠供给以及供应链不间断流动的那群人,正在加速退休。从表面看,这似乎只是自然循环,大规模退休会释放至少380万个工作岗位。但其背后隐藏着一个令人深感不安的现实:那些难以言传的隐性知识,以及通过数十年一线实践打磨出来的实操技能,正在面临随他们一同流失的风险。
尽管从人工智能、机器人到计算机视觉等技术,正在重塑工业运营方式,但从社会层面看,我们正在危险地接近失去某些能力,例如凭借声音判断电机故障,读懂模拟工程图纸,或理解一台已经有60年历史、诞生于迪斯科时代之前的老旧设备的独特脾性。
这类专业经验往往不会被集中记录,但却始终极具价值,尤其是在发生机械故障或系统级中断时。与此同时,生成式人工智能又让信息看起来触手可及。
这种矛盾是真实存在的,而且影响深远。从重工业、公用事业到供应链,摆在各行各业初级工业从业者面前的问题是:如果软件几秒钟就能够给出答案,为什么还需要花上数年时间,通过亲手实践——有时甚至经历失败——来学习?
在工业运营领域,答案其实非常简单。我们承受不起失去来之不易的知识,更不能培养一支只会使用人工智能、却不了解其所支撑系统全貌的劳动力队伍。
投资人工智能的真正机遇,在于保存维系电力供应、工厂运转和社会运作所必需的知识,并将其规模化应用。要取得成功,就必须紧跟生成式人工智能技术的发展,同时适应不断出现的宏观因素与全球挑战。这将为“人工智能与人类协同工作”(反之亦然)打开更广阔的空间,从而为未来数十年支撑全球经济的关键产业注入韧性。
人工智能的更高阶定位:并非“自动驾驶”模式
工业运转依赖的是设备可靠运行,以及管理层持续、果断地做出正确决策。但现实远没有这么简单。
在整个工业体系中,一个普遍现象是:少数经验丰富的员工,掌握着极其庞大的关键知识。他们知道哪一种振动或金属撞击声预示着故障风险,哪种权宜之计可以在物料短缺时维持生产,以及哪一张图纸准确反映了现场最新安装的硬件状态。
与此同时,许多企业在运营中,仍然依赖零散的小圈子经验、电子表格,以及需要人工整合的碎片化数据库。当某个系统宕机或某位专家退休(甚至只是请病假)时,企业几乎无法回答一些看似简单的问题,例如我们有哪些零部件?哪些资产最关键?资金究竟浪费在了哪里?
这并非只是小企业或家庭作坊的问题。制造业巨头、汽车整车厂(OEM)、车队管理公司、公用事业单位以及国防承包商,都高度依赖专业知识,也是最适合引入人工智能支持的组织类型。每家企业情况不同,但它们都面临一个人工智能可以帮助解决的核心难题:数据无处不在,却高度碎片化或相互孤立;整理这些数据,往往需要“打通”每一个系统和文件库,把相关信息汇聚到一起。若人类通过专门团队整合和梳理这些数据,可能需要数周甚至数月,而当下的人工智能只需要几分钟或几小时就能够完成对海量数据的整理。
用“决策智能”取代艰难抉择
另一个关键驱动因素在于:工业工作本身充满取舍。工厂经理、技术人员、一线维修工和工程师,几乎每天都在面对两难选择:是维修还是更换?是立即行动还是再等等?是削减成本还是降低风险?是最大化设备开机时间,还是满足可持续发展目标?这些决策牵涉数以百万计的资产,且必须在监管审查之下、常常在信息并不完整的情况下做出。人工智能的作用,是帮助人们做出更好的判断,而不是把系统调到“自动驾驶”模式,然后便置身事外。
人工智能擅长整合来自不同来源的信号,并以人类能够迅速理解的方式进行解读,比如维修记录、传感器数据、需求预测、市场环境以及环境风险等。如果运用得当,人工智能可以帮助团队进行规划、预测和优先级排序,为人类判断提供支撑。在现有技术条件下,人类和机器都不应该单打独斗。
这种辅助决策的能力,其意义早已超越了便利性或成本效率。面对电气化浪潮、数据中心扩张和全面自动化,电网、公用事业公司和制造企业正在面临前所未有的需求,因此人工智能的这种辅助能力成为一种强大的工业资产。人工智能可以更早地发现问题,为投资决策提供依据,并在安全可控的前提下延长老旧设备的使用寿命。这并不是为了自动化而自动化,而是为了确保关键系统的可靠性。
人工智能让技术岗位的劳动力更趋均衡
年轻工人(18岁至35岁)常常被批评过度依赖技术,或者在系统故障或设备需要维护时,被默认“理应”依靠技术解决问题。但现实是,他们真正想要的,是能够帮助自己安全、高效地完成有意义工作的工具。
年轻人也是最早充分意识到人工智能发展速度之快的一代。如今可用的技术,已经足以精准复现资深从业者的经验,大幅缩短学习曲线,并通过从实际工作场景中收集的近乎即时、经过验证且情境丰富的数据,弥合人才缺口。
这一点为什么重要?因为人工智能可以在不削弱岗位所需核心技能的前提下,直接打破工业岗位的入门壁垒。年轻专业人士可以利用人工智能的能力,显著减少在信息搜寻或应付碎片化系统上浪费的时间。事实上,人工智能反而让这些工作变得更具技术含量,也更有成就感,而这两点正是吸引年轻劳动力的关键。
从现场服务经理、暖通空调(HVAC)技师到工厂一线轮班工人,工业岗位维系着世界的正常运转,却常常被误解为与技术无关或是“低技能”岗位。现实恰恰相反,这些岗位需要深厚的专业知识和多元化的技能组合。在整个工业经济体系中,人工智能有望将技能培训效率提升十倍,为新一代专业工业人才打开大门。关键在于,这一过程不会牺牲质量,更不会让整个系统崩塌。
我们看到社区学院的职业教育项目报名人数有所上升,2025年较上一年增长了16%。这表明,Z世代心态开放,愿意选择蓝领工作而非传统的办公室岗位。这同样说明,人工智能不仅在发挥“均衡器”的作用,更在主动重塑和培养下一代蓝领劳动力。
将人工智能视为劳动力资产,否则将失去一切
尽管工业级人工智能才刚刚进入主流讨论视野(例如,仅2025年,数据中心合同规模就高达610亿美元,同时围绕在实体世界中全面部署人工智能而展开的GPU争夺战也愈演愈烈),但行动窗口期其实已经在迅速收窄。我估计,我们只剩下一年到两年的时间,去把数十年积累的工业知识,纳入人工智能应用以及由人工智能作为后端支撑的前沿技术平台中;否则,我们将失去这些知识,一切都将不复存在。
工业经济运行要立足于现实世界,全球无数社区依赖它提供就业、电力以及更多基础保障。面对前所未有的需求,人工智能的开发者必须推出真正实用的工具。这些工具既要尊重人类的经验,又能支持更优决策,并让复杂系统更容易理解——无论使用者是入职四周的新手,还是拥有40年经验的老员工。工业运营本身高度技术化、细致且复杂,单靠人工智能无法实现规模化运营。系统需要丰富的行业语境,也需要来自人类的合理提示,才能产出真正解决问题、并经得起时间考验的结果。
人工智能可以、也应该在工业运营中扮演权威但“辅助型”的角色,能够应用于不同行业和具体场景。
但工业界采用的创新,必须可以保留细微差别、对预测性维护的直觉以及只有多年一线实践才能形成的、针对具体事件的经验。唯有如此,才能为全球运营体系增加韧性。要在数小时或数天内完成这些工作,而不是拖到数月之后,我们既需要人工智能,也离不开人类的力量。我对未来保持乐观:人工智能不会掏空工业劳动力。事实上,大规模引入人工智能来支持更年轻一代劳动力,或许正是维持其发展的唯一途径。(财富中文网)
本文作者克里蒂·夏尔马现任IFS Nexus Black公司的首席执行官,该公司在制造业、现场服务、能源、航空航天和国防等行业部署人工智能驱动的产品。她曾经担任汤森路透(Thomson Reuters)法律科技部门的首席产品官,还曾经在赛捷集团(Sage Group)和捷孚凯(GfK)担任过人工智能相关高级职位。她的工作曾经获得英国首相颁发的“光亮之点”奖(Points of Light),并得到联合国认可;凭借在推动伦理、包容性人工智能方面的贡献,她被联合国评为“青年领袖”。
Fortune.com上发表的评论文章中表达的观点,仅代表作者本人的观点,不代表《财富》杂志的观点和立场。
译者:刘进龙
一场无声的危机,正在动摇现代社会的根基。
支撑全球经济运转的工业工人队伍,正在面临瓦解风险。负责保障电网稳定运行、工厂持续运转、公用事业可靠供给以及供应链不间断流动的那群人,正在加速退休。从表面看,这似乎只是自然循环,大规模退休会释放至少380万个工作岗位。但其背后隐藏着一个令人深感不安的现实:那些难以言传的隐性知识,以及通过数十年一线实践打磨出来的实操技能,正在面临随他们一同流失的风险。
尽管从人工智能、机器人到计算机视觉等技术,正在重塑工业运营方式,但从社会层面看,我们正在危险地接近失去某些能力,例如凭借声音判断电机故障,读懂模拟工程图纸,或理解一台已经有60年历史、诞生于迪斯科时代之前的老旧设备的独特脾性。
这类专业经验往往不会被集中记录,但却始终极具价值,尤其是在发生机械故障或系统级中断时。与此同时,生成式人工智能又让信息看起来触手可及。
这种矛盾是真实存在的,而且影响深远。从重工业、公用事业到供应链,摆在各行各业初级工业从业者面前的问题是:如果软件几秒钟就能够给出答案,为什么还需要花上数年时间,通过亲手实践——有时甚至经历失败——来学习?
在工业运营领域,答案其实非常简单。我们承受不起失去来之不易的知识,更不能培养一支只会使用人工智能、却不了解其所支撑系统全貌的劳动力队伍。
投资人工智能的真正机遇,在于保存维系电力供应、工厂运转和社会运作所必需的知识,并将其规模化应用。要取得成功,就必须紧跟生成式人工智能技术的发展,同时适应不断出现的宏观因素与全球挑战。这将为“人工智能与人类协同工作”(反之亦然)打开更广阔的空间,从而为未来数十年支撑全球经济的关键产业注入韧性。
人工智能的更高阶定位:并非“自动驾驶”模式
工业运转依赖的是设备可靠运行,以及管理层持续、果断地做出正确决策。但现实远没有这么简单。
在整个工业体系中,一个普遍现象是:少数经验丰富的员工,掌握着极其庞大的关键知识。他们知道哪一种振动或金属撞击声预示着故障风险,哪种权宜之计可以在物料短缺时维持生产,以及哪一张图纸准确反映了现场最新安装的硬件状态。
与此同时,许多企业在运营中,仍然依赖零散的小圈子经验、电子表格,以及需要人工整合的碎片化数据库。当某个系统宕机或某位专家退休(甚至只是请病假)时,企业几乎无法回答一些看似简单的问题,例如我们有哪些零部件?哪些资产最关键?资金究竟浪费在了哪里?
这并非只是小企业或家庭作坊的问题。制造业巨头、汽车整车厂(OEM)、车队管理公司、公用事业单位以及国防承包商,都高度依赖专业知识,也是最适合引入人工智能支持的组织类型。每家企业情况不同,但它们都面临一个人工智能可以帮助解决的核心难题:数据无处不在,却高度碎片化或相互孤立;整理这些数据,往往需要“打通”每一个系统和文件库,把相关信息汇聚到一起。若人类通过专门团队整合和梳理这些数据,可能需要数周甚至数月,而当下的人工智能只需要几分钟或几小时就能够完成对海量数据的整理。
用“决策智能”取代艰难抉择
另一个关键驱动因素在于:工业工作本身充满取舍。工厂经理、技术人员、一线维修工和工程师,几乎每天都在面对两难选择:是维修还是更换?是立即行动还是再等等?是削减成本还是降低风险?是最大化设备开机时间,还是满足可持续发展目标?这些决策牵涉数以百万计的资产,且必须在监管审查之下、常常在信息并不完整的情况下做出。人工智能的作用,是帮助人们做出更好的判断,而不是把系统调到“自动驾驶”模式,然后便置身事外。
人工智能擅长整合来自不同来源的信号,并以人类能够迅速理解的方式进行解读,比如维修记录、传感器数据、需求预测、市场环境以及环境风险等。如果运用得当,人工智能可以帮助团队进行规划、预测和优先级排序,为人类判断提供支撑。在现有技术条件下,人类和机器都不应该单打独斗。
这种辅助决策的能力,其意义早已超越了便利性或成本效率。面对电气化浪潮、数据中心扩张和全面自动化,电网、公用事业公司和制造企业正在面临前所未有的需求,因此人工智能的这种辅助能力成为一种强大的工业资产。人工智能可以更早地发现问题,为投资决策提供依据,并在安全可控的前提下延长老旧设备的使用寿命。这并不是为了自动化而自动化,而是为了确保关键系统的可靠性。
人工智能让技术岗位的劳动力更趋均衡
年轻工人(18岁至35岁)常常被批评过度依赖技术,或者在系统故障或设备需要维护时,被默认“理应”依靠技术解决问题。但现实是,他们真正想要的,是能够帮助自己安全、高效地完成有意义工作的工具。
年轻人也是最早充分意识到人工智能发展速度之快的一代。如今可用的技术,已经足以精准复现资深从业者的经验,大幅缩短学习曲线,并通过从实际工作场景中收集的近乎即时、经过验证且情境丰富的数据,弥合人才缺口。
这一点为什么重要?因为人工智能可以在不削弱岗位所需核心技能的前提下,直接打破工业岗位的入门壁垒。年轻专业人士可以利用人工智能的能力,显著减少在信息搜寻或应付碎片化系统上浪费的时间。事实上,人工智能反而让这些工作变得更具技术含量,也更有成就感,而这两点正是吸引年轻劳动力的关键。
从现场服务经理、暖通空调(HVAC)技师到工厂一线轮班工人,工业岗位维系着世界的正常运转,却常常被误解为与技术无关或是“低技能”岗位。现实恰恰相反,这些岗位需要深厚的专业知识和多元化的技能组合。在整个工业经济体系中,人工智能有望将技能培训效率提升十倍,为新一代专业工业人才打开大门。关键在于,这一过程不会牺牲质量,更不会让整个系统崩塌。
我们看到社区学院的职业教育项目报名人数有所上升,2025年较上一年增长了16%。这表明,Z世代心态开放,愿意选择蓝领工作而非传统的办公室岗位。这同样说明,人工智能不仅在发挥“均衡器”的作用,更在主动重塑和培养下一代蓝领劳动力。
将人工智能视为劳动力资产,否则将失去一切
尽管工业级人工智能才刚刚进入主流讨论视野(例如,仅2025年,数据中心合同规模就高达610亿美元,同时围绕在实体世界中全面部署人工智能而展开的GPU争夺战也愈演愈烈),但行动窗口期其实已经在迅速收窄。我估计,我们只剩下一年到两年的时间,去把数十年积累的工业知识,纳入人工智能应用以及由人工智能作为后端支撑的前沿技术平台中;否则,我们将失去这些知识,一切都将不复存在。
工业经济运行要立足于现实世界,全球无数社区依赖它提供就业、电力以及更多基础保障。面对前所未有的需求,人工智能的开发者必须推出真正实用的工具。这些工具既要尊重人类的经验,又能支持更优决策,并让复杂系统更容易理解——无论使用者是入职四周的新手,还是拥有40年经验的老员工。工业运营本身高度技术化、细致且复杂,单靠人工智能无法实现规模化运营。系统需要丰富的行业语境,也需要来自人类的合理提示,才能产出真正解决问题、并经得起时间考验的结果。
人工智能可以、也应该在工业运营中扮演权威但“辅助型”的角色,能够应用于不同行业和具体场景。
但工业界采用的创新,必须可以保留细微差别、对预测性维护的直觉以及只有多年一线实践才能形成的、针对具体事件的经验。唯有如此,才能为全球运营体系增加韧性。要在数小时或数天内完成这些工作,而不是拖到数月之后,我们既需要人工智能,也离不开人类的力量。我对未来保持乐观:人工智能不会掏空工业劳动力。事实上,大规模引入人工智能来支持更年轻一代劳动力,或许正是维持其发展的唯一途径。(财富中文网)
本文作者克里蒂·夏尔马现任IFS Nexus Black公司的首席执行官,该公司在制造业、现场服务、能源、航空航天和国防等行业部署人工智能驱动的产品。她曾经担任汤森路透(Thomson Reuters)法律科技部门的首席产品官,还曾经在赛捷集团(Sage Group)和捷孚凯(GfK)担任过人工智能相关高级职位。她的工作曾经获得英国首相颁发的“光亮之点”奖(Points of Light),并得到联合国认可;凭借在推动伦理、包容性人工智能方面的贡献,她被联合国评为“青年领袖”。
Fortune.com上发表的评论文章中表达的观点,仅代表作者本人的观点,不代表《财富》杂志的观点和立场。
译者:刘进龙
A silent crisis is shaking the very foundations of modern society.
The industrial workforce responsible for building the global economy is at risk of crumbling. The people charged with keeping our power grids online, factories humming, utilities reliable, and supply chains moving uninterrupted are retiring at a fast clip. Sure, this may seem like the natural cycle of things as mass retirement opens the door to at least 3.8 million jobs. But it hides a deeply troubling reality: tacit knowledge, along with practical skills refined over decades of hands-on work, is at risk of leaving with them.
While technologies from artificial intelligence to robotics to computer vision are transforming industrial operations, we’re dangerously close as a society to losing the ability to diagnose a failing motor by sound, read analog engineering drawings, or understand the quirks of a 60-year-old machine that predates Disco.
This kind of expertise is rarely written down in one place and always valuable, especially when there’s a mechanical issue or system-level disruption. Meanwhile, generative AI is making information feel instantly available.
The tension here is real and consequential. The question facing junior industrial professionals across industries, from heavy manufacturing to utilities to supply chain: If software can answer questions in seconds, why spend years learning by doing (and, in some cases, failing)?
When it comes to industrial operations, the answer is actually quite simple. We can’t afford to lose earned knowledge or train a workforce that uses AI without understanding the system it supports from soup to nuts.
The opportunity with investing in AI is to preserve the knowledge needed to keep lights on, factories humming, and society moving, and apply it at scale. Success requires keeping pace with gen AI advancements while adapting to macro factors and global challenges that come in waves. This opens the door wider for AI working with humans (and vice versa) to build resilience into essential industries powering the world’s economy for decades to come.
AI’s Elevated Role: Not On Autopilot
Industry runs on machinery and management making the right calls. Consistently. Confidently. But it’s not that simple.
Across the industrial economy, it’s common for a small group of experienced workers to serve as keepers of an outsized amount of knowledge. They know which vibration or clanking noise spells trouble, which workaround keeps production going during a shortage, and which drawing accurately reflects the latest hardware installments in the field.
At the same time, many companies still operate using a patchwork of small group expertise, spreadsheets, and fragmented databases requiring manual collation. When one system goes down or an expert retires (or, frankly, is out sick), it’s nearly impossible to answer simple questions like: what parts do we have, which assets matter most, or where is money being wasted?
These aren’t small businesses or Mom and Pop shops. Manufacturing giants, automobile OEMs, fleet management companies, utilities, and defense contractors are among the collection of expertise-dependent organizations primed for AI support. Every organization is different but they encounter the same critical problem that AI can help solve: data is everywhere, it’s fragmented or siloed, and organizing it requires plumbing every system and file repository to combine relevant information. Humans can collate and organize data collections in weeks or months with a dedicated effort. Today’s AI, meanwhile, can organize data deluges in minutes or hours.
Trade Painstaking Decisions for Decision Intelligence
The other driving factor: Industrial work is full of tradeoffs. Factory managers, technicians, floor mechanics, and engineers are constantly faced with dilemmas: fix or replace, act now or wait, cut costs or reduce risk, maximize uptime or meet sustainability goals. These decisions affect millions of assets and must be made under regulatory scrutiny, often with incomplete information. AI helps people make better decisions, not turn on autopilot and zone out.
AI is good at pulling together signals from various sources and making sense of them in a way that humans understand immediately, such as maintenance history, sensor data, demand forecasts, market conditions, and environmental risks. When used well, AI can help teams plan, predict, and prioritize. AI backstops human judgment. With the available tech, neither human or machine should be left to their own devices.
This ability to support decision-making goes beyond convenience or cost efficiency. It’s a powerful industrial asset as power grids, utilities, and manufacturers face unprecedented demands from electrification, data center growth and expansion, and full-scale automation. AI can help spot problems earlier, justify investment choices, and safely extend the life of aging equipment. That is not automation for its own sake. It is about keeping essential systems reliable.
AI: A Workforce Equalizer for Trade and Technical Work
Younger workers (18-35) are often criticized for relying too much on technology, or expected to do so when a system falters or machinery requires maintenance. In reality, they want tools that help them do meaningful work safely and efficiently.
Younger workers are also among the first groups to fully embrace that AI advances insanely fast. The tech available today is good enough to accurately reflect seasoned experience, shorten learning curves, and close talent gaps with near-instant, but verified and context-rich data gleaned from real-world work.
Why that matters: AI can demolish the barrier to entry to industrial jobs without neutering the skills required to do the job. Younger pros benefit from AI’s ability to dramatically reduce time spent mining for information or wrestling with fragmented systems. AI actually renders jobs more technical and more rewarding. Both appealing to younger workforce members.
Industrial roles from field service manager to HVAC technician to factory shift worker keep the world running, yet they are often misrepresented as tech-agnostic or low-skill. In reality, they require deep expertise and a variety of skill sets. Across the industrial economy, AI is poised to accelerate skills training ten-fold and open the door for a new generation of industrial pros to step in—and here’s the important bit—without sacrificing quality, let alone imploding the entire system.
We’re seeing vocational programs at community college enrollment numbers tick up, increasing 16% in 2025 compared to last year. This is a signal that Gen Z is open-minded and ready to take on blue collar work in favor of desk jobs. It’s also evidence that AI is not only serving as an equalizer, but actively reshaping and advancing blue collar’s next generation.
Embrace AI as a Workforce Asset, Or Lose Everything
While industrial AI is just beginning to enter mainstream conversations thanks to, for example, $61B in data center contracts in 2025 alone and a buzzy race to collect GPUs for full-scale AI deployments in the physical world, the window to act is already closing. I estimate we have 1-2 years left to capture decades of industrial knowledge in AI applications and front-edge tech platforms supported by AI on the backend, or we lose it. Everything.
The industrial economy operates in the real world, with communities around the globe relying on it for jobs, electricity, and much more. People building AI to meet unprecedented demand need to ship practical tools that respect human experience, support better decisions, and make complex systems easier to understand—whether you’ve been on the job for four weeks or 40 years. Industrial operations are deeply technical, nuanced, and complex. AI alone can’t do the work at scale. Systems need industry-rich context and informed prompts from human counterparts to produce outcomes that solve problems and stand the test of time.
AI can (and should) be applied to industrial operations in an authoritative, but supporting role across sectors and specific use cases.
But industry must adopt innovation that preserves nuance, predictive maintenance inclinations, and incident-specific experience only possible from years of hands-on work. That’s how we add resilience to global operations. To do this in hours and days, not months, we need both AI and people. I’m optimistic that AI won’t hollow out the industrial workforce. In fact, incorporating AI at scale to support a younger workforce may be the only way to sustain it.
Kriti Sharma is CEO of IFS Nexus Black, which deploys AI-powered products in manufacturing, field service, energy, aerospace, and defense industries. She was previously Chief Product Officer for LegalTech at Thomson Reuters and also previously held senior AI roles at Sage Group and GfK. Her work has been recognized by the UK Prime Minister’s Points of Light award, and the United Nations, where she was named a Young Leader for her contributions to ethical and inclusive AI.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.