,

Will AI cause unemployment? A ranking of 15 functions under impact: which work is enhanced, accelerated, or even eliminated?

Series hub / Start here|“Will AI cause unemployment?” A ranking of 15 functions under impact. I suggest you start with the overview, then go to Part B or Part C according to your function.

The enhance / accelerate / eliminate three-axis frame

  • Enhance: people still do the judgment and the relationship; AI amplifies the material, the simulation, and the insight.
  • Accelerate: the process is compressed; if you do not change the structure, you will only err faster.
  • Eliminate: tasks that can be standardized, verified, and written back are automated or largely replaced.

Series directory

Since ChatGPT broke through worldwide, the worry “will AI cause unemployment?” has kept fermenting. The real question is not whether a whole post disappears, but: does your task sit in the enhance, accelerate, or eliminate cell. Below is a directory of the 15 functions (the detailed writing has been split into the series pieces).

A directory of the 15 functions (mapped to the three axes)

For the full unfolding of the higher-risk functions, read Part B. Below, this piece keeps the functions being enhanced and the manager’s handbook (Part C).

Part C|The functions being enhanced, plus a manager’s handbook

6. Legal & Compliance

AI is making a fundamental change in the way legal and compliance work is done. In the past, companies mostly reviewed compliance after the fact — legal looked at whether a product fitted the rules only after it had been developed, or remedied a problem after it had already happened. Now, AI is moving compliance management from “after the fact” to “in real time”: through risk modeling and forecasting, a system can monitor relevant regulatory risk the moment the business is running, and warn early of a potential compliance problem. Contract review has also seen a major renewal. Traditionally, legal staff had to examine contract clauses one by one; now AI can not only help generate a first draft of a contract automatically, but can also compare a new contract quickly with a standard template or an existing version, and point out differences and potential risks at once. And in a dense, fast-changing legal environment, looking up newly issued rules by hand, line by line, can no longer keep up; AI can automatically track the latest regulatory movement in each place, compare it with the company’s current policy, and help the legal team understand in real time which changes will affect how the company runs.

In this current, AI offers legal and compliance work a powerful enhancing assistant. First, through natural-language processing and knowledge-graph technology, AI can quickly parse contract clauses, find the legal risk points hidden in them, and even generate a review summary automatically according to the company’s compliance standard, reminding legal to attend to the critical details. Second, AI tools make the tedious comparison of contract versions almost effortless — whether it is the difference among several revised versions of a contract, or the divergence of clauses between different contracts, AI can highlight them visually, and retrieve similar historical cases or precedents for reference. Beyond that, an AI platform can watch changes in the global legal and regulatory environment around the clock: when a relevant rule is updated, or a new compliance requirement appears, the system automatically pushes a reminder and attaches a summary of the change, helping the company understand its scope of effect quickly.

As AI takes over a large volume of tedious, standardized tasks, some traditional legal/compliance posts are being eliminated or reduced. Junior legal assistants, for example, used to spend a large amount of time organizing contracts and filling in regulatory filings; most of this routine work can now be finished by automation tools. Comparing contract language word by word by hand to find differences has become a thing of the past — an AI difference-comparison tool can finish in a few seconds the proofreading that would have taken several people hours. People who review compliance on a standardized process will also face a transition, because checking rules and running a checklist is more efficient and more reliable when AI does it. In this situation, the value of legal professionals stands out more clearly in the domains AI is not good at: interpreting a complex legal provision, a strategy for negotiating a contract in a distinctive situation, and professional judgment in a gray zone AI cannot decide. By handing the tedious work to AI, the legal department can concentrate its energy on higher-order risk control and strategic decision, and become a true guardian and strategic partner of the company.

7. Product Management

In the age of AI, the role of product management is recasting itself from the traditional pusher of a project into an orchestrator of tasks in “human–AI collaboration.” In the past, a product manager (PM) was mainly responsible for setting the product roadmap and coordinating cross-department resources to launch features, step by step. Now, the PM is more like a commander dispatching AI and people to work together. They no longer only arrange a schedule; they design a loop of value verification: through rapid trial and error, continually verifying whether the product design brings user value. PMs used to spend a large amount of time organizing user interviews and collecting needs; now AI tools can simulate user behavior and forecast usage situations under different design options, so that a PM can see a potential user reaction before the product has even launched. From a mere coordinator of user research, the PM is recasting themselves as a strategist who uses AI for behavioral simulation and scenario rehearsal.

In this process, AI becomes a strong enhancing assistant for the product manager. Through an intelligent-agent system, a PM can automatically assemble user feedback and product-usage data from various channels, grasp competitor movement in real time, and turn a mass of information into a basis for decision. Generative AI also helps a PM sharply shorten the experimental cycle of product development — for example, by quickly generating a product prototype, an interface design, or a line of explanatory copy, so that the PM can launch a small-scale test in a very short time, collect user views, and iterate. In addition, AI can, on the basis of early user-behavior data, forecast challenges the product may meet after launch, and even automatically produce a recommended list for the next iteration. That means a product manager can adjust product strategy more calmly on the basis of data, and does not have to rely entirely on intuition.

As AI is woven deeply into the product-development process, however, some product managers who lack a data mindset, or who only have an execution function, are facing a risk of being eliminated. A PM who traditionally has no data-analysis capability, and who decides only on experience, will find it hard to keep up with a product rhythm based on AI and data. In the same way, a large share of the work of an execution-type PM who only plays the role of “dispatching resources and tracking progress” can be replaced by automation tools — for example, a project-management system that automatically sends reminders and assembles task progress. There is also a kind of product manager who is used to stacking PowerPoint to present a plan, and who lacks empirical verification; that style will not survive in a new environment that values data and efficiency. A new generation of product-management specialists needs to master both data analysis and the skill of collaborating with AI, if they are to lead the product’s direction. They will concentrate more on seeing a user’s deeper need, setting the product vision, and working tightly with technical tools to realize that vision — becoming the critical bridge that connects user value and technical realization.

11. Corporate Strategy

The rise of AI is also renewing the way corporate strategy is set, and its rhythm. In the past, corporate strategy was often a once-a-year planning discussion at the top, with the management team applying a classic analytical frame (such as SWOT or Porter’s Five Forces) to set a blueprint for the future. Now, this “annual plan” model is being replaced by more immediate, dynamic strategic adjustment. Managers are increasingly inclined to use an experimental model: verifying a strategic hypothesis through a small-scale pilot and continuing data feedback, and correcting direction early when necessary. AI lets a company simulate various situations in a virtual environment — a sudden change in market demand, a new move by a competitor, or a change in the policy environment — and so see in advance the possible consequences of different strategic options, and reduce the blind spots of a decision. A company’s observation of an industry has also been upgraded from traditional manual work to a multi-dimensional forecasting system: AI can analyze economic data, consumer trends, and progress in technical R&D at the same time, and identify potential trends and hidden correlations faster than a person, so that strategic planning is built on a more scientific basis of anticipation.

AI offers the strategy team a rich set of enhancing tools, and gives senior decisions a stronger wing. An AI system can, for example, automatically draw a map of the competitive situation, dynamically presenting the strategic layout and movement of the main players in the market, and helping the company find a gap and an opening. At the same time, it can automatically generate a path of how a trend evolves, on the basis of a large volume of industry reports and news, so that a decision-maker can see intuitively the possible direction of a technology or a business model over the next few years. AI can even automatically produce a strategic analysis report: generating a SWOT analysis, a suggested mode of market entry, an optimization of the product portfolio, and so on, replacing a time-consuming process of making these by hand. Going further, for each strategic option, AI can quantify its risk and return in advance — through a Monte Carlo simulation or another model, estimating the financial performance of a strategy under different situations, and helping the management team fully compare the risk-return ratio of each option before deciding.

As these new capabilities appear, some traditional work of strategic analysis is being eliminated or weakened. Consultant-type analysts who used to support a decision by collecting material in large volume by hand, cutting and pasting reports, and making PowerPoint, now face direct competition from AI — a machine can, in a moment, search all public information and organize a first shape of a report. A strategy presentation that only stacks and applies a frame, and lacks a deep insight, also loses value, because an AI template already includes the standard analysis; the real difference comes from a distinctive view. In addition, drawing a static industry map (an industry landscape that only reflects the situation at one point in time) is becoming outdated; companies need more a dynamic view of information that can be updated at any time, provided by AI in real time. In the future, corporate-strategy professionals will put their energy into the domains AI cannot replace: raising a creative strategic hypothesis, comprehensively considering non-quantifiable factors such as politics and culture, and, most important, using experience and intuition to examine whether an AI recommendation is workable, and making the most considered judgment at a critical choice.

12. Corporate Communications

In the digital age, the role of corporate communications is also evolving continually because of AI’s involvement. Traditional corporate communications concentrated on raising brand exposure, increasing media coverage by issuing press releases, holding press conferences, and the like. Now, the weight has turned to shaping the brand’s narrative space in real time. That is to say, a company not only has to transmit a message; it has to keep building, across various online and offline channels, a context that favors the brand, take part in the discussion at any time, and influence the direction of public opinion. In the past, PR staff wrote press releases on a regular basis to issue news; now they attend more to how to distribute content precisely to different audiences and to build a long-term relationship of trust — issuing a message is no longer a one-way “publication,” but an interaction at the semantic level with stakeholders (customers, media, investors, and so on), to make sure the company’s values and message are understood and accepted correctly. In addition, crisis handling has moved from remedy after the fact to rehearsal beforehand: in the past, when a public-opinion crisis appeared, a statement was drafted by hand at the last moment; now AI can simulate various PR-crisis situations in advance and generate a corresponding response script, so that the company can face a sudden event more calmly.

AI tools offer the corporate-communications team several enhancing assists, and raise the efficiency and precision of communication. First, generative AI can quickly produce various kinds of PR copy — from a press release, to a senior executive’s speech, to a collaboration statement for an influencer or KOL — with AI drafting a first version for the team to polish, saving a large amount of writing time. Second, an AI-driven sentiment-monitoring system can scan social media and news movement 24 hours a day, and identify in real time a hot discussion or a potential crisis that affects the brand. When negative volume rises, the system can simulate the possible path of how the sentiment will ferment, and suggest the best communication strategy and wording, helping the PR team set the tone of a response quickly. Third, AI supports a company in communicating content across languages and cultures — automatically translating a message into local languages and adjusting the wording for localization, so that the brand’s voice stays consistent worldwide and still close to the local culture.

What follows is the elimination or weakening of some traditional corporate-communications roles. The demand for copy editors whose job is only to write drafts falls, because a first draft of the great majority of standard press releases or statements can be generated quickly by AI. The role of people who clip press coverage and monitor social media by hand is also gradually being replaced — in the past, they had to leaf through media reports every day and collect competitor information; now this information is automatically assembled and analyzed by an intelligent system, more immediately and more comprehensively. In addition, supporting staff whose only job is to make presentations and assemble communications material also face a challenge, because AI can visualize key information in real time and generate a report. In the future, corporate-communications professionals will put more of their energy into the strategic layer: shaping an inventive communication strategy, personally managing media and community relationships, and making sure every public conversation fits the company’s long-term image. In short, AI takes on a large volume of tedious execution work, and people concentrate on using creativity and judgment, guarding the brand’s reputation and building a lasting link of trust.

13. Research & Development (R&D)

The application of AI is also making a major change in research and development (R&D) work. Traditional R&D often took a linear process: a researcher set a hypothesis, then ran experiments to verify it in sequence, changing one parameter at a time to observe the result, and then deciding the next step. Now, with AI’s assistance, R&D is moving toward a model of rehearsing hypotheses in parallel: through computer simulation and machine-learning models, researchers can explore several hypothetical situations at the same time, and have AI evaluate the options with the most potential for further experiment. In the past, the tuning of many parameters in an experiment needed a researcher to keep trying; now it can be handed to AI to generate various combinations of variables automatically and look for the best solution — whether it is a formula ratio in a chemistry experiment, or the hyperparameters of a machine model, AI can, through an algorithm, quickly find the setting with the best effect. In addition, scientific knowledge used to depend mainly on a researcher’s personal accumulation and on looking up the literature by hand, and omissions were hard to avoid; with the help of AI, a research team can compare the latest academic results and patent information worldwide in time, make sure their experimental design stands on the shoulders of those who came before, and avoid walking again into a dead end others have already walked.

Various AI tools offer researchers a powerful enhancing capability, and accelerate the process of innovation. In the development of a new material or a drug, for example, AI can, according to a target characteristic, suggest a potential combination of formulas or a molecular structure, and sharply shrink the range of “trial and error.” Or AI can offer an optimization suggestion for an engineering design: from improving the shape of a product part to improving the efficiency of a software algorithm, a near-optimal scheme can be found through machine calculation. The management and analysis of experimental data also benefit from AI: the steps and results of each experiment can be automatically recorded by the system, archived in a structured way, and turned into a chart in real time for researchers to look at trends. Even in the course of an experiment, computer-vision technology can help track an observation (for example, watching a change in cells, or a subtle change in a material stress test), and offer an insight the human eye would find hard to catch in time. In the search for knowledge, researchers no longer need to spend a large amount of time leafing through the literature: AI can automatically gather related academic papers and patents, quickly organize abstracts and key findings, and help the team grasp the pulse of R&D worldwide.

As these capabilities come online, a good deal of traditional supporting work in R&D is being eliminated or reduced. The role of a technician whose job is specifically to adjust experimental parameters again and again, for example, sees its demand fall because of the introduction of automated experimental platforms and intelligent optimization algorithms — a machine can try various combinations 24 hours a day without stopping, and find the best condition faster than a person. An assistant post that helps a researcher look up literature also becomes optional; AI can finish a literature review in a short time and extract the main points. As for the work of drawing report charts by hand and organizing experimental records, it has long since been replaced by data-visualization tools and an electronic lab notebook. Researchers of the future will concentrate more on the links AI cannot replace: raising a research question that is truly original, rationally doubting and verifying a conclusion given by an AI algorithm, and turning a technical breakthrough into a concrete application. AI will become a strong assistant to the R&D team; people will be responsible for leading the direction and giving innovation its meaning.

10. Human Resources

AI is giving the substance of human-resources (HR) work a new face. In the past, the core task of an HR department was to recruit talent, review performance, and retain and develop people; it was seen as a “manager of talent.” In the age of AI, HR is gradually recasting itself as a “curator of a human–AI collaboration system.” As a large volume of repetitive work is taken over by AI, HR’s responsibility no longer only circles around people themselves. It also includes designing the mode in which people and AI work together, introducing the intelligent tools that best fit the organization, and making sure staff have the skills to work with AI. Performance management has also moved from the old formal annual KPI review toward a more dynamic and comprehensive analysis of behavior and potential: AI can track an employee’s behavioral data on various work platforms over a long period, and help judge their actual contribution, collaboration network, and growth potential, so that HR can more precisely find a star, or someone who needs help. In recruitment, the traditional process of screening résumés by hand and arranging repetitive first interviews is being replaced by intelligent matching technology — by parsing the need of a vacancy and a candidate’s material, AI can automatically filter a list of the people who fit best, and can even run a first-round interview screen with a chatbot, raising the efficiency of recruitment substantially.

AI tools also bring HR a multi-dimensional enhancing function, so that talent development and organizational care have more of a scientific basis. In the recruitment stage, for example, an AI interview assistant can score a candidate according to facial expression, tone of voice, and the content of the answer, and forecast their fitness for the role and their future potential, as a reference for the HR decision. After onboarding, each employee can have a growth roadmap personalized by AI: the system, according to their career objective and current skill, automatically recommends a training course, an on-the-job learning task, or a mentoring plan. Going further, through sentiment analysis of internal email, work chat, and similar information, AI can monitor the pulse of employee emotion and team morale in real time; once it detects a signal that pressure is too high, or that the risk of leaving is rising, it will remind HR to step in early with care. These technologies move HR management from passive to active, preventing a problem early and offering support close to a person’s need.

With these developments, some traditional HR functions are being eliminated or weakened. Recruiters who screen résumés by hand will be sharply reduced, because AI can read a mass of résumés in a short time and pick out those who meet the conditions. In the management of corporate training, the practice of tracking each employee’s course progress by hand, one by one, and teaching from a standardized PowerPoint, also fades because of the introduction of an intelligent learning platform — AI can not only remind staff to finish the training, but can also, according to the learning result, automatically give a suggestion for further study. There are also roles, such as administrative staff whose job is specifically to update personnel data in an Excel sheet or to calculate attendance bonuses, whose demand falls because of the automation functions of a human-resource management system (HRMS). Faced with the transition AI brings, HR practitioners will concentrate more on the new challenges of human–machine coexistence: shaping a sound organizational culture, setting a talent strategy for the future, attending to people’s mental health and career development, and acting as a bridge between senior management and the AI system, to make sure the application of the technology fits the organization’s values and truly enables the staff.

14. Learning & Development

Under the impetus of AI, a company’s learning and development (L&D) work is meeting a new change of model. In the past, corporate training was mainly done through a fixed course platform — staff took part in training according to a unified arrangement, and an instructor taught from material prepared in advance. Now, the method of training is gradually turning toward a flexible design of learning situations and a personalized dispatch of rhythm. That is to say, learning is no longer limited to the single form of a classroom or an online course, but is woven into everyday work situations, so that staff can obtain knowledge immediately when they need it. Companies no longer rely only on large, centralized training, but build a real-time learning-recommendation system: AI will, according to each employee’s current job need and capability gap, recommend a corresponding learning resource or a micro-course at any time, to make sure staff keep growing in the work. In the past, the effect of training often had to be evaluated after the fact through a post-class questionnaire or an exam; now, through real-time tracking and analysis of learning data, HR can grasp the progress and the effect of the training at any time, and adjust the strategy on the spot to raise the learning result.

AI tools offer a company’s L&D team a powerful enhancing assist. First, AI can automatically generate a dedicated learning map for each employee: according to personal objectives, the level of capability, and the career path, it plans the most suitable combination of courses and the order of further study, like a learning plan made to measure. AI will also filter, from a mass of internal and external knowledge bases, content that fits the employee’s need, and will push related articles, videos, or practice questions on its own. Second, at the level of delivering the training, generative AI can automatically produce simulated exam questions and interactive case exercises — in sales training, for example, AI can simulate various types of customer conversation for a learner to practice responding to, and give feedback immediately when it is over. An employee’s learning journey is recorded by the system throughout: which exercises were finished, which questions were missed repeatedly; AI will adjust the subsequent learning content accordingly, and truly teach according to the person. Finally, the management of training also becomes more efficient and more transparent: AI can automatically assemble each employee’s training result, generate a report evaluating their grasp of the knowledge and the transfer into behavior (for example, the degree of improvement in performance), and let management see clearly the return on the training investment.

As AI is involved in corporate training in a large way, some traditional content of learning-and-development work is being eliminated or recast. The role of an internal instructor who teaches standardized material from a script will gradually diminish — a method of training that repeatedly plays a PowerPoint already finds it hard to hold a learner’s interest, and is being replaced by an immersive learning experience driven by AI. Course-operations management that used to need manual work (arranging a timetable, notifying learners, collecting feedback) has also been automated; an LMS (learning management system) will automatically handle notifications and the tracking of progress. The traditional paper exam and the process of marking by hand have become history still more completely; an online test and AI scoring not only lighten the administrative load, but can also produce an analytical result immediately. In the future, L&D professionals will put their weight on building a corporate culture of lifelong learning, discovering the new skill domains the organization needs, and carefully curating various high-impact learning programs. They will collaborate more with AI, use data insight to improve the learning strategy, and make sure the company’s talent keeps evolving to meet the challenges of the future.

Closing: take up the change, and work with AI

Taken together, what AI brings is not a simple crisis of unemployment, but a full reorganization of the workplace ecology. Every functional domain is being redefined under the impact of AI: the roles that know how to collaborate with AI and produce new value will be enhanced; the parts of a process that can be automated will be accelerated by AI; and the links that lack a distinctive human value will, in the end, be eliminated. Competition in the future workplace will no longer depend on years of service, or on how familiar you are with a fixed process, but on whether you can become a “new kind of role” that can steer AI with flexibility — able to dispatch intelligent tools, connect human insight with technical capability, and create a result that the traditional model cannot reach.

For the individual, the suggested action is to take this current of transition on, on purpose. Keep learning new AI knowledge and tools, raise your own digital literacy, and treat AI as an extended partner in everyday work rather than a threat. Concentrate on building the capabilities a machine cannot replace — creative innovation, strategic thinking, empathetic communication, and cross-domain integration — so that you can play a larger role when you collaborate with AI. For the company, it should actively set a talent-development strategy for the age of AI: invest in staff’s AI-skill training, and promote the best practice of human–AI collaboration inside the organization; at the same time, re-examine the design of existing processes and posts, and be willing to adjust them to play to AI’s strengths, allocating human resources to tasks of more value. Leaders need to build a culture that encourages innovation, allow the team to try a new way of working enabled by AI, and set a clear ethical and governance frame to make sure the use of AI fits the company’s values.

We are now standing at the starting point of a historical turn. To choose to watch, or to resist, is only to be submerged by the current of change. To choose to change and to improve is to have a chance to ride AI’s east wind, and to open a new height in a career and a business. The future belongs to those who dare to work with AI: take the technology on, keep a human starting-point, and lead the way in a movement of human–machine ensemble.

The judgment, restated

My judgment is that “will I lose my job” is too coarse. First use the enhance / accelerate / eliminate axes on your task list, and only then decide whether to go into Part B or Part C.

Three things this week

  • Mark the tasks of your post with the three cells: enhance / accelerate / eliminate.
  • If the eliminate cell is the heavier one: go into Part B and align with the higher-risk approach.
  • If the enhance cell is the heavier one: rewrite “what a person does / what AI does / what we accept together,” and set an exception escalation.

Start from the Part A overview, or reply and tell us: which cell does your function sit in? You are welcome to subscribe to Future Insights Lab.

訂閱 Future Insights Lab

不定期寄送 AI × Business 短篇筆記,協助你掌握實務重點。隨時可取消訂閱。