The more reliable takeaway is this: many jobs will change because some tasks can be automated, assisted or redistributed. The ILO’s 2025 update looks at generative AI at the task level and says roughly one in four jobs is potentially transformable by GenAI. The IMF says that nearly 40% of jobs worldwide are exposed to AI-driven change.
That wording matters. “Exposed,” “affected” or “transformable” does not automatically mean “eliminated.” It means AI can alter the tasks inside a job — sometimes significantly.
The World Economic Forum’s Future of Jobs Report 2025 is based on views from more than 1,000 employers worldwide, together representing over 14 million workers. The report examines how several big forces — including technological change — may influence job growth and job decline by 2030.
That is not a personal guarantee that your profession is safe or doomed. It is a labour-market view of expected shifts in roles, tasks and skills.
The International Labour Organization, a UN agency focused on work and labour standards, describes its 2025 analysis as a refined global assessment of occupational exposure to generative AI. It combines task-level data, expert input and AI predictions to assess GenAI’s potential impact on jobs more precisely.
This is the key point for workers: two people can share the same job title and face very different levels of AI exposure. If a role contains many standardised tasks that generative AI can help with, that role is more likely to be transformed.
The International Monetary Fund describes AI as a broad force reshaping work and says nearly 40% of jobs globally are exposed to AI-driven change. It also highlights particular pressure on middle-skill roles with routine clerical or office tasks, while new skill requirements are especially visible in professional, technical and management roles.
Again, that does not mean “40% of jobs will disappear.” The stronger, better-supported reading is that AI changes which tasks are valuable and which skills workers need.
“Project manager,” “marketing manager,” “administrator,” “analyst” and “customer support specialist” sound like clear categories. In practice, the work inside those roles can vary enormously.
One marketing job may involve repeatable copy variations, basic reporting and campaign maintenance. Another may involve brand strategy, budget decisions, client negotiation and creative accountability. One administrative role may be highly standardised. Another may involve exception handling, regulatory knowledge and complex communication.
That is why the ILO’s task-based approach matters. The likely impact of generative AI depends heavily on the concrete activities people perform every week.
AI exposure tends to be higher when tasks are digital, repeatable and easy to describe. Common examples include:
This follows from the ILO’s task-level approach and the IMF’s observation that routine office work can come under particular pressure.
If AI takes on more routine work, other parts of a job can become more important. These include:
Overall, the evidence points more to changing role and skill profiles than to the simple disappearance of whole occupations. The IMF particularly emphasises new skill requirements as part of AI-driven change.
This is not a scientific scoring model. It is a practical way to apply the logic of the research: AI risk is mainly about tasks, not just job titles.
Be specific. Do not just write “I work in sales.” Write: preparing proposals, researching prospects, updating CRM records, analysing market information, coordinating with internal teams.
The more precise your list, the easier it is to see where AI could assist, accelerate or partly automate your work.
Highlight anything that is repeatable, standardised, text-heavy or data-heavy. These tasks matter because the ILO assesses generative AI at task level, while the IMF identifies routine clerical work as a pressure point.
Now highlight tasks where you carry responsibility, check outcomes, prepare decisions, negotiate with people or assess risk. These parts of work may still change because of AI, but they do not automatically disappear just because some steps become faster.
The key question is not only “What can AI do?” It is also: Can I guide AI well, evaluate its output and take responsibility for how it is used?
The IMF describes new skill requirements as a central part of AI-driven labour-market change, especially in professional, technical and management roles.
First, analyse your job at the task level. The ILO’s approach shows why differences often appear within the same occupation, not just between occupations.
Second, learn to use AI as a work tool. If AI changes routine tasks, it becomes more valuable to write clear prompts or instructions, check outputs and turn AI-generated material into usable decisions or documents.
Third, make your human contribution visible. The parts of work that remain especially important are often the ones combining context, responsibility, professional judgement, communication and decision-making.
Fourth, update your view regularly. The WEF looks at labour-market change through 2030, while the ILO and IMF describe AI as an ongoing shift in tasks and skills.
The best evidence does not answer “Will AI replace my job?” with a simple yes or no. The more defensible conclusion is that AI will change many tasks, automate or accelerate some routine work, and create new skill demands.
For your own situation, the task mix matters most. The more your work depends on repeatable digital text or data routines, the stronger the pressure for change is likely to be. The more your value comes from context, accountability, judgement and coordination, the more likely you are looking at a redesigned role rather than a straightforward replacement.