{"content_id":"okdh5nqah2","slug":"unlearning-in-the-ai-era","locale":"en","schema_type":"HowTo","category":"how_to","category_name":"How-to","title":"Unlearning in the AI Era: How to Let Go of Old Assumptions and Learn Again","summary":"Unlearning is not about forcing yourself to forget knowledge, but about identifying assumptions and work methods that are no longer valid and reducing their influence. This explains how to test outdated assumptions, conduct small experiments, and replace them with new criteria for judgment.","author":{"name":"Injoys Editorial Team","url":"https://injoys.com/ko/about"},"key_points":["Unlearning is not about erasing memories, but about reassessing the conditions under which existing knowledge still applies.","The half-life of job skills is not a single standardized statistic, so a specific number of years should not be treated as a universal fact.","Effective transitions occur through a cycle of unlearning, relearning, and new learning.","Rather than immediately discarding existing methods, use small experiments to compare the effectiveness and risks of new approaches.","Even when AI generates information, responsibility for setting goals, judging context, verification, and outcomes remains with people and organizations."],"content_markdown":"If the feeling of falling behind persists even though you keep learning new technologies, the problem may be less about how much you learn and more about outdated assumptions that remain in your mind. In an environment where AI and automation are rapidly changing work processes, you need not only the ability to add new knowledge but also **unlearning**, the practice of reassessing the scope in which existing knowledge applies.\n\n## What Is Unlearning?\n\nUnlearning does not mean completely erasing what you have learned or rendering past experience worthless. It is **the process of examining whether knowledge, habits, or assumptions that were once effective remain valid in the current situation and preventing outdated elements from continuing to dominate decision-making**.\n\nFor example, in the past, it may have been reasonable for people to write every report from scratch. In an environment where generative AI is available, the drafting process can be changed, but responsibility for fact-checking and approval does not automatically disappear. Unlearning does not mean discarding an entire existing process. It means distinguishing again between what should be automated and what people must continue to be responsible for.\n\n### The Difference Between Unlearning and Forgetting\n\n| Category | Meaning | Result |\n|---|---|---|\n| Forgetting | Losing memories or information regardless of intent | Existing knowledge may not be available when needed |\n| Discarding | Deciding not to use specific information or procedures | A gap in operations may arise if there is no alternative |\n| Unlearning | Reassessing the assumptions and conditions under which existing knowledge applies | Valid elements can be preserved while outdated ones are replaced |\n| Relearning | Learning core concepts or skills again in a changed environment | Existing experience is reorganized according to new standards |\n\n## Why Unlearning Matters in the AI Era\n\n### The Combination of Skills Required for Jobs Is Changing Rapidly\n\nIn the World Economic Forum’s *Future of Jobs Report 2025*, the employers surveyed expected approximately 39% of workers’ existing skill sets to change or become outdated between 2025 and 2030. This does not mean that all skills will disappear at the same rate. Rather, it means that the composition and relative importance of the skills required for many jobs are likely to change significantly.\n\nThe period over which the market value of a job skill falls by half is often described as a “skill half-life.” However, this figure varies greatly by occupation, skill, industry, and measurement method, and there is no common international standard metric. Therefore, rather than accepting statements such as “the half-life of every skill is 5 years” or “the half-life of AI skills is 2 years” as universal facts, it is more accurate to treat them as metaphors explaining the direction of change or as estimates with limited scope.\n\n### New Tools Change the Assumptions Behind Existing Processes\n\nWork processes are generally created based on the costs and technological constraints of their time. Typical examples include research procedures developed when searching took a long time, approval systems designed around in-person work, and production processes based on the assumption that people write every draft.\n\nWhen AI supports search, summarization, translation, coding, and document drafting, some of these assumptions change. This does not mean that all existing processes become invalid. Some conditions must be maintained even when tools change, including privacy protection, copyright, security, fact-checking, and final accountability.\n\n### Knowing More and Exercising Better Judgment Are Different\n\nAI can quickly present many options, but it does not decide which option best serves an organization’s objectives or determine the acceptable level of risk. Nor is responsibility for judging whether an output is factual, useful to users, and compliant with laws and organizational policies automatically transferred.\n\nCapabilities in the AI era therefore cannot be explained simply by the amount of information a person possesses. The following abilities are also necessary.\n\n- The ability to define problems and goals accurately\n- The ability to verify sources and evidence\n- The ability to distinguish important information from unnecessary information\n- The ability to reflect the context of the organization and its users\n- The ability to explain results and take responsibility for them\n\n## The Cycle of Unlearning, Relearning, and New Learning\n\nUnlearning is not an activity that ends on its own. After letting go of existing assumptions, you must understand the essentials again and connect new methods to actual work.\n\n| Stage | Core Question | Main Activities | Output |\n|---|---|---|---|\n| Unlearning | What is no longer true or essential? | Identifying assumptions, reviewing conditions of application, analyzing exceptions | A list of items to discontinue, retain, or verify |\n| Relearning | What objectives and principles have not changed? | Redefining goals, checking the latest evidence, revising standards | New criteria for judgment and operating principles |\n| New learning | What can be created with new tools and conditions? | Experimentation, measurement, feedback, process design | A validated new way of working |\n\nThese three stages form a recurring cycle rather than a straight line. Even new methods become subject to reassessment when the environment changes.\n\n## 6 Steps for Putting Unlearning into Practice\n\n### 1. Look for Signals of Change\n\nIf the following patterns recur, it may be necessary to examine the assumptions behind existing methods.\n\n- More time is being invested, but the quality of results is not improving.\n- New tools have been introduced, but the work stages remain exactly the same.\n- Regulations, technology, or user requirements have changed.\n- Team members cannot explain the purpose of a procedure and simply follow convention.\n- The same mistakes or rework recur.\n\n### 2. Write Hidden Assumptions as Statements\n\nTurn the beliefs hidden behind procedures into specific statements. Avoid expressions such as “This is how it has always been done,” because they cannot be verified.\n\nExamples include:\n\n- A good report must be written directly by a person from beginning to end.\n- You must attend meetings for a long time to understand the context of the work.\n- Managers must know every practical task better than their team members.\n- The longer a document is, the more thoroughly it has been reviewed.\n\n### 3. Distinguish Facts, Assumptions, and Preferences\n\nClassify each statement into one of the following three categories.\n\n- **Fact:** Something that can be confirmed with current evidence\n- **Assumption:** Something that is true only under certain conditions or has not yet been verified\n- **Preference:** A way of doing things desired by an individual or organization\n\nTreating preferences as facts makes unlearning difficult. Conversely, mistaking legal obligations or safety standards for mere conventions and removing them can be dangerous.\n\n### 4. Check the Conditions of Application and Period of Validity\n\nIdentify when the existing method was created, the problem it was intended to solve, and the tools available at the time. Then ask whether the same conditions still apply today.\n\n- What problem was this rule created to prevent?\n- Does that problem still exist?\n- Have the technology or cost structure changed?\n- Do the same regulatory, security, or quality standards still apply?\n- What is the greatest risk of discontinuing this method?\n\n### 5. Run Small Experiments Instead of Replacing Everything\n\nUnlearning is not indiscriminate disposal. Compare the existing method with a new one within a reversible scope.\n\nFor example, before eliminating weekly reports entirely, one team could reduce the number of reporting items by half and compare decision-making time, errors, rework, and user satisfaction. When introducing generative AI, it is also safer to begin by testing the drafting of documents that can be made public while maintaining separate control procedures for sensitive data and final approval.\n\n### 6. Record the New Standards and Reasons for Discarding the Old Ones\n\nBased on the results of the experiment, retain, revise, suspend, or discard the existing method. Do not record only what changed; also document the following.\n\n- The purpose of the previous method\n- Assumptions that are no longer valid\n- Principles and controls that must be maintained\n- Conditions under which the new method is valid\n- The reassessment date and the person responsible\n\nThis record prevents the new method from becoming yet another unverified convention.\n\n## Examples of Workplace Applications\n\n### Report Writing\n\n- **Existing assumption:** Quality is high only when the writer personally reads all materials and writes every sentence from scratch.\n- **Unlearning:** Examine whether there is always an inevitable relationship between direct authorship and accuracy.\n- **Relearning:** Redefine the essence of a report as a document with clear evidence that is useful for decision-making.\n- **New learning:** Use AI to create the structure or draft, while people compare it against the original sources, verify figures, and provide final approval.\n\n### Running Meetings\n\n- **Existing assumption:** Everyone involved must attend meetings in real time.\n- **Unlearning:** Distinguish whether information sharing and joint decision-making require the same type of meeting.\n- **Relearning:** Clarify whether the purpose of a meeting is to convey information, gather opinions, or make a decision.\n- **New learning:** Move information sharing to asynchronous documents and use meetings only for agenda items requiring debate and decisions.\n\n### The Manager’s Role\n\n- **Existing assumption:** Managers must know every answer and give direct instructions.\n- **Unlearning:** Determine whether this approach slows decision-making in an organization where expertise is distributed.\n- **Relearning:** Redefine the manager’s role as setting goals, allocating resources, resolving conflicts, and managing accountability.\n- **New learning:** Delegate practical judgments to those responsible and create a structure for reviewing important assumptions and risks.\n\n## What Must Not Be Discarded\n\nRapid change does not justify discarding every principle. Even when introducing new tools, the following must be carefully preserved or strengthened.\n\n- Compliance with laws and regulations\n- Protection of personal and confidential information\n- Essential verification for safety and quality\n- Source verification and records management\n- Principles of accessibility and nondiscrimination\n- Final decision-making authority and accountability\n\nAI outputs in particular may contain plausible but incorrect information, bias, or the risk of exposing sensitive information. Eliminating verification and accountability procedures in the name of efficiency is closer to a failure of control than to unlearning.\n\n## Common Reasons Unlearning Fails\n\n### Rejecting All Past Experience\n\nExisting methods were created for reasons that addressed problems at the time. Erasing that history may cause previously resolved risks to reappear. Instead of rejecting people or their experience, examine the conditions under which that experience applies separately.\n\n### Treating a Trendy Tool as the Answer\n\nSomething is not necessarily better simply because it is new. If adopting a tool becomes the goal itself, it may only increase cost and complexity. Measurable goals such as quality, speed, risk, and user value must be defined first.\n\n### Discarding Existing Methods Without Preparing Alternatives\n\nIf there is only discontinuation without relearning, a gap in operations will arise. Identify the function performed by the procedure being discarded and establish new standards to replace that function.\n\n### Ignoring the Psychological Cost\n\nExpertise developed over many years is closely connected to identity. A request to change an existing method may sound like a denial of an individual’s abilities. It is more effective to explain which conditions have changed rather than what was wrong.\n\n## An Unlearning Checklist You Can Start Today\n\nYou can begin simply by choosing and answering one of the questions below.\n\n1. What work has remained unchanged even though the environment changed during the past 1 year?\n2. Are there procedures explained with words such as “originally,” “obviously,” or “always”?\n3. What was the original problem that the procedure was intended to solve?\n4. Do the same problem and constraints still exist?\n5. What are the core principles that must be maintained, and what formats can be changed?\n6. What is the smallest change that can be tested while limiting risk?\n7. What metrics will be used to judge success and failure?\n\n## Conclusion\n\nUnlearning is not an argument for no longer learning. It is **a method of reassessing the conditions under which existing knowledge remains valid in order to create room for new learning to work**. What matters is not discarding as much as possible, but developing the ability to judge what to retain and why, and what to replace and on what evidence.\n\nThere is no need to change every method today. You can begin by writing down one assumption behind something you thought “has always been done this way” and using a small experiment to determine whether that assumption remains valid.","content_html":"\u003cp\u003eIf the feeling of falling behind persists even though you keep learning new technologies, the problem may be less about how much you learn and more about outdated assumptions that remain in your mind. In an environment where AI and automation are rapidly changing work processes, you need not only the ability to add new knowledge but also \u003cstrong\u003eunlearning\u003c/strong\u003e, the practice of reassessing the scope in which existing knowledge applies.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#what-is-unlearning\" class=\"anchor\" id=\"what-is-unlearning\"\u003e\u003c/a\u003eWhat Is Unlearning?\u003c/h2\u003e\n\u003cp\u003eUnlearning does not mean completely erasing what you have learned or rendering past experience worthless. It is \u003cstrong\u003ethe process of examining whether knowledge, habits, or assumptions that were once effective remain valid in the current situation and preventing outdated elements from continuing to dominate decision-making\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFor example, in the past, it may have been reasonable for people to write every report from scratch. In an environment where generative AI is available, the drafting process can be changed, but responsibility for fact-checking and approval does not automatically disappear. Unlearning does not mean discarding an entire existing process. It means distinguishing again between what should be automated and what people must continue to be responsible for.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#the-difference-between-unlearning-and-forgetting\" class=\"anchor\" id=\"the-difference-between-unlearning-and-forgetting\"\u003e\u003c/a\u003eThe Difference Between Unlearning and Forgetting\u003c/h3\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eCategory\u003c/th\u003e\n\u003cth\u003eMeaning\u003c/th\u003e\n\u003cth\u003eResult\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Category\"\u003eForgetting\u003c/td\u003e\n\u003ctd data-label=\"Meaning\"\u003eLosing memories or information regardless of intent\u003c/td\u003e\n\u003ctd data-label=\"Result\"\u003eExisting knowledge may not be available when needed\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Category\"\u003eDiscarding\u003c/td\u003e\n\u003ctd data-label=\"Meaning\"\u003eDeciding not to use specific information or procedures\u003c/td\u003e\n\u003ctd data-label=\"Result\"\u003eA gap in operations may arise if there is no alternative\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Category\"\u003eUnlearning\u003c/td\u003e\n\u003ctd data-label=\"Meaning\"\u003eReassessing the assumptions and conditions under which existing knowledge applies\u003c/td\u003e\n\u003ctd data-label=\"Result\"\u003eValid elements can be preserved while outdated ones are replaced\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Category\"\u003eRelearning\u003c/td\u003e\n\u003ctd data-label=\"Meaning\"\u003eLearning core concepts or skills again in a changed environment\u003c/td\u003e\n\u003ctd data-label=\"Result\"\u003eExisting experience is reorganized according to new standards\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003ch2\u003e\n\u003ca href=\"#why-unlearning-matters-in-the-ai-era\" class=\"anchor\" id=\"why-unlearning-matters-in-the-ai-era\"\u003e\u003c/a\u003eWhy Unlearning Matters in the AI Era\u003c/h2\u003e\n\u003ch3\u003e\n\u003ca href=\"#the-combination-of-skills-required-for-jobs-is-changing-rapidly\" class=\"anchor\" id=\"the-combination-of-skills-required-for-jobs-is-changing-rapidly\"\u003e\u003c/a\u003eThe Combination of Skills Required for Jobs Is Changing Rapidly\u003c/h3\u003e\n\u003cp\u003eIn the World Economic Forum’s \u003cem\u003eFuture of Jobs Report 2025\u003c/em\u003e, the employers surveyed expected approximately 39% of workers’ existing skill sets to change or become outdated between 2025 and 2030. This does not mean that all skills will disappear at the same rate. Rather, it means that the composition and relative importance of the skills required for many jobs are likely to change significantly.\u003c/p\u003e\n\u003cp\u003eThe period over which the market value of a job skill falls by half is often described as a “skill half-life.” However, this figure varies greatly by occupation, skill, industry, and measurement method, and there is no common international standard metric. Therefore, rather than accepting statements such as “the half-life of every skill is 5 years” or “the half-life of AI skills is 2 years” as universal facts, it is more accurate to treat them as metaphors explaining the direction of change or as estimates with limited scope.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#new-tools-change-the-assumptions-behind-existing-processes\" class=\"anchor\" id=\"new-tools-change-the-assumptions-behind-existing-processes\"\u003e\u003c/a\u003eNew Tools Change the Assumptions Behind Existing Processes\u003c/h3\u003e\n\u003cp\u003eWork processes are generally created based on the costs and technological constraints of their time. Typical examples include research procedures developed when searching took a long time, approval systems designed around in-person work, and production processes based on the assumption that people write every draft.\u003c/p\u003e\n\u003cp\u003eWhen AI supports search, summarization, translation, coding, and document drafting, some of these assumptions change. This does not mean that all existing processes become invalid. Some conditions must be maintained even when tools change, including privacy protection, copyright, security, fact-checking, and final accountability.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#knowing-more-and-exercising-better-judgment-are-different\" class=\"anchor\" id=\"knowing-more-and-exercising-better-judgment-are-different\"\u003e\u003c/a\u003eKnowing More and Exercising Better Judgment Are Different\u003c/h3\u003e\n\u003cp\u003eAI can quickly present many options, but it does not decide which option best serves an organization’s objectives or determine the acceptable level of risk. Nor is responsibility for judging whether an output is factual, useful to users, and compliant with laws and organizational policies automatically transferred.\u003c/p\u003e\n\u003cp\u003eCapabilities in the AI era therefore cannot be explained simply by the amount of information a person possesses. The following abilities are also necessary.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eThe ability to define problems and goals accurately\u003c/li\u003e\n\u003cli\u003eThe ability to verify sources and evidence\u003c/li\u003e\n\u003cli\u003eThe ability to distinguish important information from unnecessary information\u003c/li\u003e\n\u003cli\u003eThe ability to reflect the context of the organization and its users\u003c/li\u003e\n\u003cli\u003eThe ability to explain results and take responsibility for them\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\n\u003ca href=\"#the-cycle-of-unlearning-relearning-and-new-learning\" class=\"anchor\" id=\"the-cycle-of-unlearning-relearning-and-new-learning\"\u003e\u003c/a\u003eThe Cycle of Unlearning, Relearning, and New Learning\u003c/h2\u003e\n\u003cp\u003eUnlearning is not an activity that ends on its own. After letting go of existing assumptions, you must understand the essentials again and connect new methods to actual work.\u003c/p\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eStage\u003c/th\u003e\n\u003cth\u003eCore Question\u003c/th\u003e\n\u003cth\u003eMain Activities\u003c/th\u003e\n\u003cth\u003eOutput\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Stage\"\u003eUnlearning\u003c/td\u003e\n\u003ctd data-label=\"Core Question\"\u003eWhat is no longer true or essential?\u003c/td\u003e\n\u003ctd data-label=\"Main Activities\"\u003eIdentifying assumptions, reviewing conditions of application, analyzing exceptions\u003c/td\u003e\n\u003ctd data-label=\"Output\"\u003eA list of items to discontinue, retain, or verify\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Stage\"\u003eRelearning\u003c/td\u003e\n\u003ctd data-label=\"Core Question\"\u003eWhat objectives and principles have not changed?\u003c/td\u003e\n\u003ctd data-label=\"Main Activities\"\u003eRedefining goals, checking the latest evidence, revising standards\u003c/td\u003e\n\u003ctd data-label=\"Output\"\u003eNew criteria for judgment and operating principles\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Stage\"\u003eNew learning\u003c/td\u003e\n\u003ctd data-label=\"Core Question\"\u003eWhat can be created with new tools and conditions?\u003c/td\u003e\n\u003ctd data-label=\"Main Activities\"\u003eExperimentation, measurement, feedback, process design\u003c/td\u003e\n\u003ctd data-label=\"Output\"\u003eA validated new way of working\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003cp\u003eThese three stages form a recurring cycle rather than a straight line. Even new methods become subject to reassessment when the environment changes.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#6-steps-for-putting-unlearning-into-practice\" class=\"anchor\" id=\"6-steps-for-putting-unlearning-into-practice\"\u003e\u003c/a\u003e6 Steps for Putting Unlearning into Practice\u003c/h2\u003e\n\u003ch3\u003e\n\u003ca href=\"#1-look-for-signals-of-change\" class=\"anchor\" id=\"1-look-for-signals-of-change\"\u003e\u003c/a\u003e1. Look for Signals of Change\u003c/h3\u003e\n\u003cp\u003eIf the following patterns recur, it may be necessary to examine the assumptions behind existing methods.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eMore time is being invested, but the quality of results is not improving.\u003c/li\u003e\n\u003cli\u003eNew tools have been introduced, but the work stages remain exactly the same.\u003c/li\u003e\n\u003cli\u003eRegulations, technology, or user requirements have changed.\u003c/li\u003e\n\u003cli\u003eTeam members cannot explain the purpose of a procedure and simply follow convention.\u003c/li\u003e\n\u003cli\u003eThe same mistakes or rework recur.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\n\u003ca href=\"#2-write-hidden-assumptions-as-statements\" class=\"anchor\" id=\"2-write-hidden-assumptions-as-statements\"\u003e\u003c/a\u003e2. Write Hidden Assumptions as Statements\u003c/h3\u003e\n\u003cp\u003eTurn the beliefs hidden behind procedures into specific statements. Avoid expressions such as “This is how it has always been done,” because they cannot be verified.\u003c/p\u003e\n\u003cp\u003eExamples include:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eA good report must be written directly by a person from beginning to end.\u003c/li\u003e\n\u003cli\u003eYou must attend meetings for a long time to understand the context of the work.\u003c/li\u003e\n\u003cli\u003eManagers must know every practical task better than their team members.\u003c/li\u003e\n\u003cli\u003eThe longer a document is, the more thoroughly it has been reviewed.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\n\u003ca href=\"#3-distinguish-facts-assumptions-and-preferences\" class=\"anchor\" id=\"3-distinguish-facts-assumptions-and-preferences\"\u003e\u003c/a\u003e3. Distinguish Facts, Assumptions, and Preferences\u003c/h3\u003e\n\u003cp\u003eClassify each statement into one of the following three categories.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eFact:\u003c/strong\u003e Something that can be confirmed with current evidence\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eAssumption:\u003c/strong\u003e Something that is true only under certain conditions or has not yet been verified\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePreference:\u003c/strong\u003e A way of doing things desired by an individual or organization\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eTreating preferences as facts makes unlearning difficult. Conversely, mistaking legal obligations or safety standards for mere conventions and removing them can be dangerous.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#4-check-the-conditions-of-application-and-period-of-validity\" class=\"anchor\" id=\"4-check-the-conditions-of-application-and-period-of-validity\"\u003e\u003c/a\u003e4. Check the Conditions of Application and Period of Validity\u003c/h3\u003e\n\u003cp\u003eIdentify when the existing method was created, the problem it was intended to solve, and the tools available at the time. Then ask whether the same conditions still apply today.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eWhat problem was this rule created to prevent?\u003c/li\u003e\n\u003cli\u003eDoes that problem still exist?\u003c/li\u003e\n\u003cli\u003eHave the technology or cost structure changed?\u003c/li\u003e\n\u003cli\u003eDo the same regulatory, security, or quality standards still apply?\u003c/li\u003e\n\u003cli\u003eWhat is the greatest risk of discontinuing this method?\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\n\u003ca href=\"#5-run-small-experiments-instead-of-replacing-everything\" class=\"anchor\" id=\"5-run-small-experiments-instead-of-replacing-everything\"\u003e\u003c/a\u003e5. Run Small Experiments Instead of Replacing Everything\u003c/h3\u003e\n\u003cp\u003eUnlearning is not indiscriminate disposal. Compare the existing method with a new one within a reversible scope.\u003c/p\u003e\n\u003cp\u003eFor example, before eliminating weekly reports entirely, one team could reduce the number of reporting items by half and compare decision-making time, errors, rework, and user satisfaction. When introducing generative AI, it is also safer to begin by testing the drafting of documents that can be made public while maintaining separate control procedures for sensitive data and final approval.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#6-record-the-new-standards-and-reasons-for-discarding-the-old-ones\" class=\"anchor\" id=\"6-record-the-new-standards-and-reasons-for-discarding-the-old-ones\"\u003e\u003c/a\u003e6. Record the New Standards and Reasons for Discarding the Old Ones\u003c/h3\u003e\n\u003cp\u003eBased on the results of the experiment, retain, revise, suspend, or discard the existing method. Do not record only what changed; also document the following.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eThe purpose of the previous method\u003c/li\u003e\n\u003cli\u003eAssumptions that are no longer valid\u003c/li\u003e\n\u003cli\u003ePrinciples and controls that must be maintained\u003c/li\u003e\n\u003cli\u003eConditions under which the new method is valid\u003c/li\u003e\n\u003cli\u003eThe reassessment date and the person responsible\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThis record prevents the new method from becoming yet another unverified convention.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#examples-of-workplace-applications\" class=\"anchor\" id=\"examples-of-workplace-applications\"\u003e\u003c/a\u003eExamples of Workplace Applications\u003c/h2\u003e\n\u003ch3\u003e\n\u003ca href=\"#report-writing\" class=\"anchor\" id=\"report-writing\"\u003e\u003c/a\u003eReport Writing\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eExisting assumption:\u003c/strong\u003e Quality is high only when the writer personally reads all materials and writes every sentence from scratch.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eUnlearning:\u003c/strong\u003e Examine whether there is always an inevitable relationship between direct authorship and accuracy.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRelearning:\u003c/strong\u003e Redefine the essence of a report as a document with clear evidence that is useful for decision-making.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNew learning:\u003c/strong\u003e Use AI to create the structure or draft, while people compare it against the original sources, verify figures, and provide final approval.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\n\u003ca href=\"#running-meetings\" class=\"anchor\" id=\"running-meetings\"\u003e\u003c/a\u003eRunning Meetings\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eExisting assumption:\u003c/strong\u003e Everyone involved must attend meetings in real time.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eUnlearning:\u003c/strong\u003e Distinguish whether information sharing and joint decision-making require the same type of meeting.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRelearning:\u003c/strong\u003e Clarify whether the purpose of a meeting is to convey information, gather opinions, or make a decision.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNew learning:\u003c/strong\u003e Move information sharing to asynchronous documents and use meetings only for agenda items requiring debate and decisions.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\n\u003ca href=\"#the-managers-role\" class=\"anchor\" id=\"the-managers-role\"\u003e\u003c/a\u003eThe Manager’s Role\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eExisting assumption:\u003c/strong\u003e Managers must know every answer and give direct instructions.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eUnlearning:\u003c/strong\u003e Determine whether this approach slows decision-making in an organization where expertise is distributed.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRelearning:\u003c/strong\u003e Redefine the manager’s role as setting goals, allocating resources, resolving conflicts, and managing accountability.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNew learning:\u003c/strong\u003e Delegate practical judgments to those responsible and create a structure for reviewing important assumptions and risks.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\n\u003ca href=\"#what-must-not-be-discarded\" class=\"anchor\" id=\"what-must-not-be-discarded\"\u003e\u003c/a\u003eWhat Must Not Be Discarded\u003c/h2\u003e\n\u003cp\u003eRapid change does not justify discarding every principle. Even when introducing new tools, the following must be carefully preserved or strengthened.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eCompliance with laws and regulations\u003c/li\u003e\n\u003cli\u003eProtection of personal and confidential information\u003c/li\u003e\n\u003cli\u003eEssential verification for safety and quality\u003c/li\u003e\n\u003cli\u003eSource verification and records management\u003c/li\u003e\n\u003cli\u003ePrinciples of accessibility and nondiscrimination\u003c/li\u003e\n\u003cli\u003eFinal decision-making authority and accountability\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eAI outputs in particular may contain plausible but incorrect information, bias, or the risk of exposing sensitive information. Eliminating verification and accountability procedures in the name of efficiency is closer to a failure of control than to unlearning.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#common-reasons-unlearning-fails\" class=\"anchor\" id=\"common-reasons-unlearning-fails\"\u003e\u003c/a\u003eCommon Reasons Unlearning Fails\u003c/h2\u003e\n\u003ch3\u003e\n\u003ca href=\"#rejecting-all-past-experience\" class=\"anchor\" id=\"rejecting-all-past-experience\"\u003e\u003c/a\u003eRejecting All Past Experience\u003c/h3\u003e\n\u003cp\u003eExisting methods were created for reasons that addressed problems at the time. Erasing that history may cause previously resolved risks to reappear. Instead of rejecting people or their experience, examine the conditions under which that experience applies separately.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#treating-a-trendy-tool-as-the-answer\" class=\"anchor\" id=\"treating-a-trendy-tool-as-the-answer\"\u003e\u003c/a\u003eTreating a Trendy Tool as the Answer\u003c/h3\u003e\n\u003cp\u003eSomething is not necessarily better simply because it is new. If adopting a tool becomes the goal itself, it may only increase cost and complexity. Measurable goals such as quality, speed, risk, and user value must be defined first.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#discarding-existing-methods-without-preparing-alternatives\" class=\"anchor\" id=\"discarding-existing-methods-without-preparing-alternatives\"\u003e\u003c/a\u003eDiscarding Existing Methods Without Preparing Alternatives\u003c/h3\u003e\n\u003cp\u003eIf there is only discontinuation without relearning, a gap in operations will arise. Identify the function performed by the procedure being discarded and establish new standards to replace that function.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#ignoring-the-psychological-cost\" class=\"anchor\" id=\"ignoring-the-psychological-cost\"\u003e\u003c/a\u003eIgnoring the Psychological Cost\u003c/h3\u003e\n\u003cp\u003eExpertise developed over many years is closely connected to identity. A request to change an existing method may sound like a denial of an individual’s abilities. It is more effective to explain which conditions have changed rather than what was wrong.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#an-unlearning-checklist-you-can-start-today\" class=\"anchor\" id=\"an-unlearning-checklist-you-can-start-today\"\u003e\u003c/a\u003eAn Unlearning Checklist You Can Start Today\u003c/h2\u003e\n\u003cp\u003eYou can begin simply by choosing and answering one of the questions below.\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eWhat work has remained unchanged even though the environment changed during the past 1 year?\u003c/li\u003e\n\u003cli\u003eAre there procedures explained with words such as “originally,” “obviously,” or “always”?\u003c/li\u003e\n\u003cli\u003eWhat was the original problem that the procedure was intended to solve?\u003c/li\u003e\n\u003cli\u003eDo the same problem and constraints still exist?\u003c/li\u003e\n\u003cli\u003eWhat are the core principles that must be maintained, and what formats can be changed?\u003c/li\u003e\n\u003cli\u003eWhat is the smallest change that can be tested while limiting risk?\u003c/li\u003e\n\u003cli\u003eWhat metrics will be used to judge success and failure?\u003c/li\u003e\n\u003c/ol\u003e\n\u003ch2\u003e\n\u003ca href=\"#conclusion\" class=\"anchor\" id=\"conclusion\"\u003e\u003c/a\u003eConclusion\u003c/h2\u003e\n\u003cp\u003eUnlearning is not an argument for no longer learning. It is \u003cstrong\u003ea method of reassessing the conditions under which existing knowledge remains valid in order to create room for new learning to work\u003c/strong\u003e. What matters is not discarding as much as possible, but developing the ability to judge what to retain and why, and what to replace and on what evidence.\u003c/p\u003e\n\u003cp\u003eThere is no need to change every method today. You can begin by writing down one assumption behind something you thought “has always been done this way” and using a small experiment to determine whether that assumption remains valid.\u003c/p\u003e\n","tags":["Unlearning","Relearning","AI Era","Job Competencies","Learning Methods"],"faqs":[{"question":"Is unlearning about forgetting what you have learned?","answer":"No. Unlearning is not about erasing memories, but about reexamining the assumptions and conditions under which existing knowledge or habits were valid and applicable. Principles that remain valid are retained, while only outdated elements are given less weight or replaced with new approaches."},{"question":"How are unlearning and relearning different?","answer":"Unlearning is the stage of moving away from assumptions that are no longer valid, while relearning is the stage of understanding concepts and skills anew to fit a changed environment. Generally, unlearning creates room, relearning establishes new standards, and then these are connected to practical experiments at work."},{"question":"What is the actual half-life of job skills in years?","answer":"There is no standard half-life that applies to all job skills. The rate at which the value of skills changes varies by industry, occupation, region, and measurement criteria. A specific number of years may be an estimate from an individual study or analogy, so its source and definition should be checked together."},{"question":"Does unlearning mean ignoring existing experience?","answer":"No. Existing experience is valuable information that shows why a particular approach was developed. Instead of discarding experience, unlearning distinguishes between the conditions under which that experience still applies and those under which it no longer does."},{"question":"What is the safest way to begin unlearning in an organization?","answer":"It is safest to start with small-scale, reversible experiments. Compare the quality, time, cost, errors, and risks of the existing and new approaches, and maintain mandatory controls related to legal, security, and safety requirements throughout the experiment."},{"question":"As AI advances, will human judgment no longer be necessary?","answer":"AI can generate more options and drafts, but it does not independently determine an organization's goals, acceptable risks, user value, or ultimate accountability. Responsibility for fact-checking, contextual judgment, approval, and outcomes must be assigned to a clearly designated person or organization."},{"question":"How can the outcomes of unlearning be measured?","answer":"Rather than measuring how much new knowledge has been learned, it is more useful to measure reductions in rework, decision-making time, error rates, quality, user satisfaction, and whether risks materialize. Baseline values must be established before the change, and the same metrics should be compared after a set period."},{"question":"What should I do first to start unlearning today?","answer":"Choose one procedure at work that is explained as, “This is how we’ve always done it.” Write down the problem the procedure was intended to solve, the constraints that existed when it was created, and the risks that still remain, then design the smallest possible experiment to change it."}],"sources":[{"url":"https://www.weforum.org/publications/the-future-of-jobs-report-2025/","title":"World Economic Forum, The Future of Jobs Report 2025","type":"data_point"},{"url":"https://hbr.org/2016/11/why-the-problem-with-learning-is-unlearning","title":"Harvard Business Review, Why the Problem with Learning Is Unlearning","type":"source"},{"url":"https://books.google.com/books?q=%22By+instructing+students+how+to+learn%2C+unlearn+and+relearn%22","title":"Alvin Toffler, Future Shock","type":"expert_quote"},{"url":"https://www.nist.gov/itl/ai-risk-management-framework","title":"NIST AI Risk Management Framework","type":"source"}],"images":[{"id":336,"url":"https://injoys.com/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6MzkxMywicHVyIjoiYmxvYl9pZCJ9fQ==--895d9fef4ea5fbfb5ab9a739efd6c892766fd90c/ai-4fd2b3f4.webp","is_representative":true,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"깨진 톱니바퀴, 나침반, 윤리 상징, AI 큐브를 순환 화살표로 연결한 일러스트","caption":"낡은 지식을 버리고 방향과 윤리를 점검하며 AI를 다시 배우는 과정을 나타낸다.","description":null},"en":{"alt":"Illustration linking broken gears, a compass, ethics symbols, and an AI cube with circular arrows","caption":"The cycle represents discarding old assumptions and relearning AI with direction and ethics.","description":null},"ja":{"alt":"壊れた歯車、コンパス、倫理の象徴、AIキューブを循環する矢印で結んだイラスト","caption":"古い前提を捨て、方向性と倫理を確かめながらAIを学び直す過程を表している。","description":null},"es":{"alt":"Ilustración de engranajes rotos, una brújula, símbolos éticos y un cubo de IA unidos por flechas","caption":"El ciclo representa desechar viejas premisas y reaprender la IA con orientación y ética.","description":null},"id":{"alt":"Ilustrasi roda gigi rusak, kompas, simbol etika, dan kubus AI yang dihubungkan panah melingkar","caption":"Siklus ini menggambarkan melepas asumsi lama dan mempelajari kembali AI dengan arah serta etika.","description":null},"pt":{"alt":"Ilustração de engrenagens quebradas, bússola, símbolos éticos e cubo de IA ligados por setas","caption":"O ciclo representa abandonar velhas premissas e reaprender IA com direção e ética.","description":null},"zh-hant":{"alt":"破裂齒輪、指南針、倫理符號與AI方塊由循環箭頭連結的插畫","caption":"循環呈現拋開舊有前提，並兼顧方向與倫理重新學習AI的過程。","description":null},"de":{"alt":"Illustration mit kaputten Zahnrädern, Kompass, Ethiksymbolen und KI-Würfel im Pfeilkreislauf","caption":"Der Kreislauf steht dafür, alte Annahmen abzulegen und KI mit Orientierung und Ethik neu zu lernen.","description":null}}},{"id":337,"url":"https://injoys.com/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6MzkxOSwicHVyIjoiYmxvYl9pZCJ9fQ==--c2124f79e9b74bcb94be864024ff33010e5e4e50/ai-a5e4790e.webp","is_representative":false,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"사람들이 기존 지식과 AI 기반의 새 학습 경로를 비교하는 과정도 일러스트","caption":"낡은 지식을 걷어내고 AI 시대에 맞는 새로운 판단 체계를 구축하는 과정을 보여준다.","description":null},"en":{"alt":"People compare old knowledge and AI-driven learning paths in a workflow diagram","caption":"The illustration shows a shift from outdated assumptions to AI-enabled ways of learning and deciding.","description":null},"ja":{"alt":"人々が従来の知識とAIを活用した新しい学習経路を比較する工程図","caption":"古い前提を手放し、AI時代に合う学び方と判断基準を築く過程を表している。","description":null},"es":{"alt":"Personas comparan el conocimiento antiguo con una nueva ruta de aprendizaje basada en IA","caption":"La ilustración muestra el paso de viejos supuestos a nuevas formas de aprender y decidir con IA.","description":null},"id":{"alt":"Orang-orang membandingkan pengetahuan lama dan jalur pembelajaran baru berbasis AI","caption":"Ilustrasi ini menunjukkan peralihan dari asumsi lama menuju cara belajar dan mengambil keputusan dengan AI.","description":null},"pt":{"alt":"Pessoas comparam conhecimentos antigos e uma nova jornada de aprendizagem com IA","caption":"A ilustração mostra a transição de antigas premissas para novas formas de aprender e decidir com IA.","description":null},"zh-hant":{"alt":"人們透過流程圖比較舊有知識與人工智慧驅動的新學習路徑","caption":"圖中呈現拋開過時假設並建立人工智慧時代新學習與決策方式的過程。","description":null},"de":{"alt":"Menschen vergleichen altes Wissen mit einem neuen KI-gestützten Lernpfad","caption":"Die Illustration zeigt den Wechsel von überholten Annahmen zu neuen Lern- und Entscheidungswegen mit KI.","description":null}}}],"published_at":"2026-07-29T05:29:15+09:00","updated_at":"2026-07-29T05:29:15+09:00","license":"cc_by","translation_status":"reviewed","available_locales":["ko","en","ja","es"],"data_locales":["ko","en","ja","es","id","pt","zh-hant","de"],"url":"https://injoys.com/en/articles/unlearning-in-the-ai-era"}