Robin Lu
学习 观察 10 分钟阅读

AI 时代下,岗位正在被重新拆解与重组

1. AI 对当前岗位的解构与重组

在我看来,AI 在未来一段时间内对就业市场最深刻的影响,并不是简单地“取代多少岗位”,而是更隐蔽但更彻底的——把原本完整的岗位拆解成更细的任务单元,再重新组合出新的角色形态。

在此前的工业 4.0 阶段,技术进步更多改变的是“工具形态”,而对“岗位结构”的影响相对间接。 传统工业体系中,人之所以深度嵌入生产流程,是因为机器要运转,离不开人的实时操作与监控;即便出现了机器与机器(M2M)的协同,整体流程的启动、异常处理以及很多关键调整,仍然主要依赖于人。

而在这一波 AI 浪潮之下,人的位置正悄悄发生变化。 越来越多的场景里,人不再是亲自执行每一个步骤的“操作员”,而是在更高一层,去:

  • 设定目标和约束条件
  • 决定在何时、以何种方式启动某项任务
  • 判断 AI 给出的结果是否足够好、是否可以被采用

AI 系统开始接手大量信息输入、初步处理、草稿生成等环节。 在这样的前提下,一份传统意义上的“岗位”,往往可以被拆解为多个可以交给 AI 处理的子任务,而人负责的是更靠前的“问题定义”和更靠后的“结果判断”。

这会带来一个明显的结构性变化:

  • 在“任务层面”,大量重复、规则清晰的工作逐步转移给 AI
  • 在“岗位层面”,人的工作被重新围绕“理解问题、判断场景、协调多个任务”来组合

从这个意义上讲,人不再只是庞大工业机器中的一个零件,而更像是负责“决定做什么、做到什么程度算完成”的那一方。 当然,这种转变在不同企业和行业中的推进速度并不一致,在相当多的一线岗位里,具体操作和决策判断仍然是长期共存的。

因此,对于个体能力的要求,也在从“会做某个具体动作”,转向“是否真正理解这项任务在整个业务流程中的意义与边界”。 换句话说,未来岗位对人的偏好,将更侧重于:对行业与任务的理解深度、对问题的拆解能力,以及与 AI 协同完成工作的能力,而不仅仅是机械执行某一项技能的熟练度。


2. AI 对当前工作体系的替代性体现在哪里?

在谈论“替代性”的时候,人们很容易陷入一种简单的二元对立——仿佛 AI 会直接把人的岗位“拿走”。 但如果细看现实正在发生的变化,会发现 AI 更深层的影响,其实体现在:知识门槛的改变、企业用人标准的调整,以及人机之间任务分工方式的重构。

过去几十年里,可获取的信息总量在急剧膨胀。 从搜索引擎、在线课程,到现在随时可用的 AI 助手,公共知识的获取成本被不断压低。 这意味着,对于大量通识和应用型知识来说,“知道去哪儿找到它、能多快理解并用起来”,正在变得比“是否早就把它牢牢记在脑子里”更重要。

在传统的教育体系下,学历在很大程度上被视为“知识存量”的一个粗略信号。 随着高等教育的普及,拿到学历的人越来越多,单纯以“文凭多少”来区分人才的效果自然开始减弱。 企业在招聘上的重心,也逐步从单一看重学历,转向更加关注实际能力、项目经验以及可验证的工作成果。

AI 的出现,又明显加速了这一趋势。 对于许多知识工作者来说,AI 可以:

  • 帮助快速检索并整合公开资料
  • 生成报告、文案、分析的初稿
  • 以更低门槛的方式讲解复杂概念

在这样的环境下,个体之间在“脑中储存了多少公共知识”上的差异,正变得不再像过去那么关键。 真正拉开差距的是:

  • 谁能在 AI 的辅助下,更快理解和吸收新知识
  • 谁能更好地按需求把问题拆成具体可执行的步骤
  • 谁能把零散的信息重新组合成对业务有价值的解决方案或输出

从这个角度重新审视“AI 的替代性”,我们会发现,风险的高低,与其说由“行业名称”决定,不如说由“任务特征”决定:

  • 那些技能单一、任务高度重复、几乎只需按照固定指令执行的岗位,往往暴露程度更高
  • 那些对整体流程的理解要求不高、个体不需要做太多判断与反思的工作,更容易被抽象成可被模型学习的模式

相对而言,需要现场体力劳动、频繁面对面服务、复杂人际沟通的岗位,在短期内更难被完全用 AI 替掉。 但这并不意味着它们完全“无风险”——更现实的情况是,它们当中的部分环节,会被自动化或数字化技术改造,人类在其中的角色,会从单纯体力执行,转向更多设备协同、现场协调与服务体验提升。

因此,与其说 AI 直接“大规模消灭岗位”,不如说它正在“重排任务”: 一些人原本负责的基础性工作,会被工具接手; 而那些能持续学习、善于理解业务、懂得如何利用 AI 完成更复杂目标的人,将在这一轮重排中获得更大的空间。


3. AI 时代下,风向的变化:焦虑、机会与个体价值

如果把视角从单一岗位拉远一些,会看到一条更长的趋势线: 一方面,在数字化和 AI 的共同作用下,人类工业和组织的整体效率在持续提升; 另一方面,正因为越来越多的例行性任务可以被工具接管,真正属于“人的那一部分价值”,反而被凸显出来。

那些能够把 AI 视作工具,而不是威胁的人,通常会表现出更低的焦虑感。 他们会刻意把重复性工作交给 AI,腾出时间和精力,用在决策、创新、沟通、跨界整合这些更需要人类判断和创造力的地方。

随着效率的提升,试错成本被摊薄,商业世界对更多细分市场和更具体的个性化需求,展现出比过去更高的兴趣。 在诸多小众领域和高度个性化的场景中,完全标准化、规模化的方案往往难以覆盖所有细节,人的审美、共情、判断与创造,就有了发光的空间。

从这个意义上讲,AI 时代并不只是简单地“把所有人推向同一个模板”。 它更像是放大镜:

  • 对那些被动等待系统安排、固守旧有技能、不愿学习新工具的人来说,技术变革可能放大的是不确定感和被替代风险
  • 而对那些愿意主动理解业务、持续打磨思维模型、学会指挥 AI 完成更复杂目标的人来说,技术则会放大他们原本就存在的差异化优势

1. How AI Decomposes and Reassembles Jobs

In my view, the most profound near‑term impact of AI on the job market is not simply “how many jobs it replaces”, but something more subtle and far‑reaching: it breaks down traditional roles into smaller task units and then recombines them into new types of positions.

In the earlier phase of Industry 4.0, technological progress mainly changed the form of tools, while its influence on the internal structure of jobs was relatively indirect. In traditional industrial systems, humans were deeply embedded in production lines because machines could not function without real‑time human operation and supervision. Even when machine‑to‑machine (M2M) collaboration emerged, the initiation of workflows, the handling of exceptions, and many key adjustments still largely depended on humans.

In the current wave of AI, the position of humans in the workflow is quietly changing. In more and more scenarios, people are no longer the ones personally executing every single step. Instead, they operate at a higher level to:

  • define goals and constraints
  • decide when and how to initiate a task
  • judge whether the AI‑generated result is good enough to adopt

AI systems are starting to take over large portions of information input, preliminary processing, and draft generation. Under these conditions, a traditionally defined “job” can often be decomposed into multiple subtasks that can be handed off to AI, while humans remain responsible for earlier “problem definition” and later “result evaluation”.

This leads to a clear structural shift:

  • At the task level, many repetitive and rule‑based activities are progressively delegated to AI.
  • At the role level, human work is recombined around understanding problems, interpreting contexts, and coordinating multiple tasks.

In this sense, people are no longer just interchangeable components in a large industrial machine. They become more like the ones who decide “what should be done” and “to what standard it should be done”. Of course, this transition is progressing at different speeds across industries and organizations, and in many frontline positions, hands‑on operations and decision‑making responsibilities will coexist for a long time.

Therefore, the capability required of individuals is shifting from “being able to perform a specific action” toward “truly understanding the meaning and boundaries of that task within the overall business process”. In other words, future roles will favor those who have deep understanding of their industry and tasks, strong problem‑decomposition skills, and the ability to collaborate effectively with AI, rather than those who only excel at mechanically executing a single narrow skill.


2. Where Does AI’s Substitution Effect Really Show Up?

When people talk about “substitution”, they often fall into a simple binary: as if AI will directly “take away” human jobs. However, if we look more closely at what is actually changing, we see that AI’s deeper impact lies in how it reshapes knowledge barriers, hiring criteria, and the division of labor between humans and machines.

Over the past few decades, the total volume of accessible information has exploded. From search engines and online courses to today’s always‑available AI assistants, the cost of acquiring public knowledge has been continuously pushed down. For a large amount of general and applied knowledge, “knowing where to find it and how quickly you can understand and use it” is becoming more important than “whether you memorized it long ago”.

In traditional education systems, formal degrees were widely used as a rough signal of how much knowledge a person had accumulated. As higher education expanded and more people obtained degrees, the effectiveness of using diplomas alone to distinguish talent naturally declined. Companies have gradually shifted their hiring focus from pure degree requirements toward practical capabilities, project experience, and verifiable work outcomes.

The emergence of AI has clearly accelerated this shift. For many knowledge workers, AI can:

  • help quickly retrieve and synthesize public information
  • generate first drafts of reports, copy, or analyses
  • explain complex concepts at a lower learning threshold

In this environment, differences between individuals in terms of “how much public knowledge they store in their heads” are no longer as decisive as before. What truly differentiates people is:

  • who can, with AI’s assistance, understand and absorb new knowledge more quickly
  • who can better break problems down into concrete, executable steps
  • who can recombine scattered information into solutions or outputs that create real business value

From this angle, when we reassess “AI’s substitution effect”, we find that risk levels are determined less by industry labels and more by task characteristics:

  • Roles with very narrow skills and highly repetitive tasks that mostly follow fixed instructions tend to have higher exposure.
  • Jobs that do not require deep understanding of the overall process and ask little of an individual’s judgment or reflection are easier to abstract into patterns that models can learn.

By contrast, roles that require on‑site physical labor, frequent face‑to‑face service, and complex human interaction are, at least in the short term, more difficult to fully replace with AI. But this does not mean they are entirely “risk‑free”. A more realistic picture is that certain components of these jobs will be automated or digitized, and human roles will shift from pure physical execution toward more equipment collaboration, on‑site coordination, and experience enhancement.

Therefore, rather than saying AI is directly “wiping out jobs at scale”, it is more accurate to say that AI is reallocating tasks. Some basic work that people used to do will be taken over by tools. Meanwhile, those who keep learning, understand the business deeply, and know how to leverage AI to accomplish more complex goals will gain more room to grow in this reallocation process.


3. Changing Winds in the AI Era: Anxiety, Opportunity, and Individual Value

If we zoom out from individual roles and look at the broader landscape, a longer‑term pattern emerges. On one hand, overall industrial and organizational efficiency keeps improving under the combined influence of digitalization and AI. On the other hand, precisely because more routine tasks can be handled by tools, the part of value that truly belongs to humans is becoming more visible.

Those who can treat AI as a tool rather than a threat usually experience lower levels of anxiety. They deliberately hand repetitive tasks over to AI and free up time and mental energy for areas that demand human judgment and creativity, such as decision‑making, innovation, communication, and cross‑domain integration.

As efficiency improves and the cost of experimentation drops, the business world is showing greater interest in niche markets and more specific personalized needs. In many niche and highly individualized scenarios, purely standardized and large‑scale solutions often struggle to cover every detail. This creates space for human aesthetics, empathy, judgment, and creativity to shine.

In this sense, the AI era is not simply pushing everyone into the same mold. It functions more like a magnifying glass:

  • For those who passively wait for the system to arrange everything, cling to old skills, and resist learning new tools, technological change may amplify uncertainty and the risk of being displaced.
  • For those who actively seek to understand their business, continually refine their mental models, and learn to orchestrate AI to accomplish more complex goals, technology will amplify the differentiated advantages they already possess.