AI has accelerated work, but it does not automatically improve the way people learn and grow. Behind rising productivity, new challenges are emerging, including weaker motivation and gaps in expertise. Organizations must now design not only for speed, but also for learning, judgment, and long-term growth.'

***

As artificial intelligence entered the workplace, the first thing to change was speed. Tasks that once required hours of searching for and organizing information can now be completed in minutes, while report drafts, meeting summaries, market analyses, and customer response messages can also be produced in a short period of time. Work that previously required a certain level of experience and training can now be performed to a considerable standard even by beginners with the help of artificial intelligence.

This change is undeniably attractive. Repetitive work has decreased, and the burden of getting started on a task has also been lowered. Small organizations can produce more materials and content than before, while departments that lack specialized personnel can still generate results of a consistent quality. As the cost of turning ideas into actual outputs has fallen, the speed at which new ideas can be tested has also increased.

However, the fact that work has become faster does not mean that people are learning more effectively. While artificial intelligence produces drafts and answers on their behalf, people may fail to understand fully why those results were produced. Performance may improve while interest in and responsibility for work weaken, and the quality of outputs may rise without people developing the ability to produce the same results independently.

The central question of the age of artificial intelligence is not limited to how much work can be automated. More important challenges now involve how increased speed can be connected to human growth and how employees can preserve a sense of meaning and accomplishment in an automated environment. Rather than simply introducing technology, redesigning the processes through which motivation, learning, and expertise are formed is becoming a new source of organizational competitiveness.

The most direct change that artificial intelligence has brought to organizations is the compression of time. The time required to write reports, conduct research, classify data, create program code, and produce images has been dramatically reduced. Workflows have also shifted away from a model in which people create everything from the beginning toward one in which artificial intelligence presents a basic framework and people revise it.

Employees have been able to reduce the time spent on repetitive work and concentrate on more complex problems. The cost of turning ideas into actual results has also declined. It has become easier to convert a concept that previously existed only in someone?™s mind into a document, compare multiple versions, and try again after failure. Artificial intelligence has moved the starting point of work forward and lowered the threshold for experimentation.

Yet as work has accelerated, the performance standards demanded by organizations have risen as well. At a time when one person produced one report per day, that level of output was recognized as appropriate performance. Now, however, there is an expectation that the same person should be able to produce several outputs in the same amount of time by using artificial intelligence. Technology that was initially introduced to reduce workloads has gradually begun to serve as a standard for demanding more work.

When response times become shorter, customers expect faster replies. When the time required to write documents decreases, managers demand more analyses and reports. The time saved is increasingly likely to be filled with new assignments rather than used for rest or learning. Artificial intelligence may have reduced the time required for work, but it has not necessarily increased the sense of breathing room experienced by employees.

In this process, the ability to produce work quickly has begun to be regarded as an important competitive advantage. Time spent deeply understanding a problem, discussing it with colleagues, or critically reviewing information can easily appear slow and inefficient. As immediate answers and rapid drafts become commonplace, careful and deliberate thinking may be treated as a sign of insufficient competitiveness.

Faster results, however, do not always mean better results. Sentences produced by artificial intelligence may sound natural and appear well organized, yet they can contain factual errors or overlook essential context. The more convincing the expression appears, the more difficult it may be to detect the error. Accurately reviewing an output requires knowledge, experience, and the judgment needed to understand the relevant context.

As artificial intelligence increases speed, the human capacity for verification becomes even more important. When organizations measure performance only through output volume and time saved, employees focus on submitting results quickly. By contrast, when accuracy, depth of understanding, and long-term usefulness are evaluated together, artificial intelligence can move beyond being a tool for speed competition and become a tool that supports better judgment.

People want to reduce work that is difficult and burdensome. It is also true that workloads decrease when repetitive document organization, simple data entry, and the preparation of materials in fixed formats are automated. However, making work easier and increasing motivation toward work are two different matters.

The satisfaction people experience in their work is not created by the final output alone. A sense of accomplishment develops through the process of understanding a problem, finding a method, and resolving it through trial and error. Work also gains greater meaning when people feel that their abilities have improved or when they solve a difficult task on their own. When the work process becomes excessively compressed, these experiences may disappear along with it.

A person who spends a long time researching information, constructing an argument, and completing a report is likely to develop a sense of ownership over the result. By contrast, someone whose role is limited to briefly revising a draft generated by artificial intelligence may finish the report without clearly feeling what they personally contributed. The result has been produced more quickly, but the sense of accomplishment may have declined.

When artificial intelligence proposes ideas and even writes the sentences, employees may begin to see themselves as reviewers rather than creators. Reviewing is also important, but attachment to the work weakens when people cannot feel how much their judgment has influenced the result. Those with limited work experience may believe that artificial intelligence produces better results than they do and may stop attempting new tasks altogether.

Motivation is also closely connected to autonomy. People participate more actively in work when they feel that they can choose their methods and influence the results. However, when an organization excessively standardizes the use of artificial intelligence and forces every employee to follow the same tools and procedures, work can once again become a mechanical process. Automation may reduce simple work while also reducing human discretion.

When artificial intelligence provides answers quickly, there is less reason to spend a long time struggling with difficult problems. People may develop the habit of asking artificial intelligence for help whenever they encounter even a small obstacle instead of searching for a solution themselves. This may appear efficient in the immediate term, but over time it can weaken the ability to endure problems and maintain concentration.

An appropriate degree of difficulty at work can stimulate growth. Removing every inconvenience does not necessarily create a good working environment. Unnecessary repetition and formal procedures should be reduced, but the processes of thinking, choosing, and taking responsibility should remain. To preserve employee motivation, organizations need to distinguish between work that can be delegated to artificial intelligence and work that people should perform themselves.

Organizations should not stop at instructing employees to use artificial intelligence. They need to clarify which problems should be entrusted to technology, which judgments should remain human, and how human contributions are reflected in the final output. When employees understand their role not simply as correcting results but as setting direction and establishing standards, artificial intelligence can become a tool that expands autonomy rather than one that weakens motivation.

Expertise does not simply mean possessing a large amount of knowledge. It also includes the ability to judge which knowledge should be used in a particular situation and to respond appropriately when an unexpected problem occurs. Such abilities are not developed merely by listening to explanations or checking correct answers. They are gradually formed through accumulated repetition, mistakes, feedback, and revision.

In the past, beginners started with simple tasks and gradually took responsibility for more complex work. They repeatedly searched for information themselves, wrote drafts, received criticism from senior colleagues, and revised their work. Although this process may have appeared slow and inefficient, it was essential training through which they internalized the principles of the work and the standards required for judgment.

Artificial intelligence has greatly shortened this stage. Even beginners can use artificial intelligence to produce documents and analyses that appear to have been created by experienced professionals. The visible difference in performance has narrowed, and organizations may easily conclude that new employees can be assigned directly to practical work. Yet the ability to produce an output and the ability to understand it are not the same.

Someone may use code generated by artificial intelligence to complete a program but remain unable to identify the cause when an error occurs. A report may be written quickly, while the writer remains unable to determine whether its core assumptions are incorrect. A natural-sounding translation may be produced without the user noticing that an important meaning has changed. The quality of the output has improved, but the ability to solve problems independently has not developed sufficiently.

This situation creates an illusion of expertise. Employees themselves may believe that they possess greater ability than they actually do, while managers may judge their capabilities to be sufficient by looking only at the finished output. The problem may remain hidden under ordinary conditions, but limitations become apparent when artificial intelligence produces an incorrect answer or an unexpected situation arises.

A more serious problem is that the path through which future experts are developed may weaken. The people responsible for advanced judgment within organizations are usually those who accumulated extensive experience with basic work and trial and error in the past. If all entry-level work is automated, the next generation will have fewer opportunities to build expertise. Current productivity may rise, while the organization gradually faces a shortage of internal experts over the long term.

In the past, repetitive work served as a starting point for learning. Through simple information organization and draft writing, employees understood the structure of the work and learned standards from small mistakes. If artificial intelligence performs all of these tasks, organizations must deliberately provide alternative learning experiences. Experiences lost through automation need to be reconstructed through education and training.

This does not mean that beginners should avoid using artificial intelligence. What matters is the process of analyzing and explaining the results rather than accepting them as they are. Employees should be required to examine why a particular answer was produced, which assumptions were used, what alternative options exist, and what kinds of errors may be possible. The process of using artificial intelligence itself should be designed as an opportunity for learning.

Methods of evaluation must also change. Instead of looking only at the final output, organizations should ask what questions were posed, which parts of the artificial intelligence response were revised, and what evidence supported the final judgment. Only then can simple copying be distinguished from genuine understanding. An expert is not merely someone who finds an answer quickly, but someone who understands the limitations of an answer and can adapt it to the situation.

Learning in the age of artificial intelligence must move from transmitting knowledge toward training judgment. The ability to remember and repeat necessary information remains important, but the ability to assess the reliability of information and apply it appropriately within a particular context has become more valuable. The more artificial intelligence provides answers, the more people must take responsibility for creating good questions and verifying the answers.

Work processes should not remain limited to a structure in which artificial intelligence creates something first and a person merely approves it. Employees should first define the problem and develop their own hypotheses before comparing them with the answers produced by artificial intelligence. When every answer is delegated from the beginning, human thinking can easily become passive. By contrast, when people make an initial judgment and then use technology, they can learn by comparing their own thinking with the results produced by artificial intelligence.

When writing a report, for example, a person can first decide the central argument and structure instead of immediately generating the entire draft. In data analysis, employees can think about which variables should be examined and what results they expect before asking for a conclusion. Artificial intelligence should be used not as a tool that replaces thinking, but as one that expands and verifies it.

The process of explanation is also important. Asking employees to explain artificial intelligence-generated results to colleagues in their own words makes it possible to assess the depth of their understanding. Knowledge that cannot be explained is difficult to use in real work situations. People must be able to describe why a conclusion is valid and under what conditions the result might change before they can take responsibility for it.

Feedback within organizations must also change. In the past, much time was spent correcting expressions and formatting errors. If artificial intelligence can handle these aspects, managers can focus on higher-level questions. Feedback should concentrate on whether the problem was defined appropriately, whether important stakeholders were overlooked, whether the standards for judgment were clear, and whether long-term risks were considered.

Training programs must also be connected to actual work. When education focuses only on the functions of tools and how to use them, employees merely learn to operate new technology. Training should also cover situations in which artificial intelligence should not be used, methods for verifying results, and ways of handling confidential information and copyright issues. Principles for responsible use are more important than technical functions alone.

Within a certain scope, training without artificial intelligence is also necessary. To maintain core capabilities, people occasionally need experience analyzing information and writing drafts by themselves. There is no need to return every task to manual methods, but deliberate practice is necessary to prevent essential abilities from weakening.

It is also unwise to divide the roles of people and artificial intelligence in a fixed manner. Their roles can vary depending on the difficulty and risk of the task and the employee?™s level of experience. Technology can take on a greater share of simple and repetitive work, but people must remain deeply involved in important decisions and unfamiliar problems. Beginners need guaranteed opportunities to learn, while experienced employees need an environment in which they can concentrate on complex judgment.

The success or failure of artificial intelligence adoption may depend less on which technology an organization chooses than on what it recognizes as performance. When only output volume and cost reduction are rewarded, employees will delegate as much work as possible to artificial intelligence. Those who spend time on verification and learning may instead be judged as slow and inefficient.

Organizations must examine the accumulation of capability alongside productivity. They need to assess whether employees have become able to solve new problems, detect errors made by artificial intelligence, explain their own judgments, and transfer knowledge to colleagues. A balance is required between short-term performance and long-term growth.

The role of managers is also changing. In the past, their main responsibilities centered on assigning work and reviewing results. Now, designing the structure of collaboration between people and artificial intelligence has become increasingly important. Managers must decide which tasks should be automated and which experiences should remain with employees. Increasing speed while protecting opportunities for learning is becoming a new managerial capability.

The psychological responses of employees also require careful attention. The introduction of artificial intelligence creates both expectation and anxiety. Some people quickly learn to use new tools, while others worry that their experience and expertise may no longer be valued. If organizations do not adequately explain the purpose of adopting the technology and how roles will change, employees may view artificial intelligence not as a tool for growth but as a tool for surveillance and replacement.

How organizations use the time saved through artificial intelligence is equally important. When all saved time is filled with additional work, employees find it difficult to experience the benefits of the technology. Some of that time should be available for learning, experimentation, collaboration, customer understanding, and long-term assignments. Only then can productivity improvements lead to an expansion of organizational capability rather than merely an increase in work intensity.

The structure of collaboration between experts and beginners must also be strengthened. Even when artificial intelligence generates drafts quickly, the contextual knowledge and experience of experts remain important. Experts should move beyond simply correcting errors and take responsibility for explaining and transferring the standards behind their judgments. Beginners can learn by comparing the results produced by artificial intelligence with the judgments made by experienced colleagues.

Organizations also need spaces where new methods of using artificial intelligence can be tested in low-risk work. If perfect performance is demanded from the outset, employees will use only familiar and safe functions or conceal errors. Sharing failures as well as successes enables the entire organization to learn more quickly.

The view that artificial intelligence is merely a technology for eliminating human shortcomings must also be approached with caution. People are slow and make mistakes, but through that process they learn to understand context, accept responsibility, and build expertise. If all of these processes are removed in the name of efficiency, organizations may gain faster outputs while losing the capacity for deep judgment.

Competitiveness in the age of artificial intelligence will not belong only to organizations that automate the largest amount of work. Organizations that use the speed provided by technology while preserving human motivation and expertise will be able to build stronger capabilities over a longer period of time. Technology may reduce the time required for work, but whether that time is connected to learning and growth depends on the choices made by the organization.

In a rapidly transformed working environment, the ability people need is not simply the ability to work faster. It is the ability to define problems accurately, question the answers produced by artificial intelligence, take responsibility for judgment, and share experience with colleagues. Organizations cannot manage speed alone. They must also examine why employees work, what they are learning, and which capabilities they are leaving to the next generation.

Productivity can be expressed in numbers, but motivation and expertise are formed over a long period of time. Organizations must deliberately protect the process of human growth rather than being drawn only toward short-term efficiency gains. As the range of work performed by artificial intelligence expands, the experience of thinking, judging, and taking responsibility directly becomes more valuable.

Artificial intelligence is neither simply a tool that eliminates human work nor an all-purpose technology that solves every problem. Its meaning changes according to which tasks are reduced and which experiences are preserved. When increased speed can be converted into room for learning and growth, artificial intelligence can become not a technology that weakens expertise, but a foundation that enables a higher level of expertise.

Harvard Business School, March 2026, Navigating the Jagged Technological Frontier: Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality

Microsoft Research, April 2025, The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects from a Survey of Knowledge Workers

OECD, June 2026, AI and Skills: What We Know So Far

OECD, June 2025, The Effects of Generative AI on Productivity, Innovation and Entrepreneurship

World Economic Forum, January 2025, The Future of Jobs Report 2025