What LinkedIn’s Work Change Report 2025 Actually Shows

What does the 70% headline measure, and what does it leave open?

If a headline says seven in ten skills will change, it is easy to read it as seven in ten duties being automated. That is not what LinkedIn claims. Its January 2025 report says it expects 70% of skills used in most jobs to change by 2030, with AI emerging as a catalyst. LinkedIn’s press release labels this an expectation, so it should be read as a forecast, not a count of changes already observed.

The report’s retrospective 140% figure is different: since 2022, the pace at which LinkedIn members add new skills to profiles increased by that amount. Its method tracks skill groups members add across occupations on the platform. A profile addition may represent learning, a change in how someone describes existing work, or a new capability; by itself it does not establish a changed task, proficiency, or cause. LinkedIn’s dataset offers a broad platform view, but profile activity is not a census of workers or a direct inventory of their duties.

The report explains how it measured profile skill breadth and several other indicators, but the opened methodology does not give a reproducible calculation for the 70% forecast. That limits independent assessment of how the estimate was derived. It does not make the forecast meaningless; it does mean the number cannot support a precise claim about any one occupation or person. It certainly does not mean that 70% of your tasks will be replaced by AI.

Sources: Work Change Report: AI Is Coming to Work; Work Change Report: Skills for jobs set to change by 70% by 2030

Which signals show change already happening?

The report brings together several kinds of signal, and they answer different questions. LinkedIn describes more members adding AI skills, skills appearing in paid job posts, and employers or executives reporting priorities and experiences. Its global C-suite research surveyed 1,991 executives in nine countries, all at businesses with at least 1,000 employees; its business survey covered more than 2,500 businesses in five countries. Those findings describe the people and organizations surveyed, not every employer’s current practice.

LinkedIn’s separate 2025 Skills on the Rise ranking makes the distinction especially clear. It combines three measures: members adding a skill to profiles, hiring success among members with the skill, and increased presence in paid job postings. Those are useful signals of platform activity and demand within LinkedIn’s data. They are not a forecast of future skill change, proof that a skill caused someone to be hired, or evidence that organizations have implemented AI across all relevant workflows.

For your own decision, look for corroboration close to your work: which tasks recur, whether your employer permits a tool for them, what your manager says the team needs, and what relevant postings in your location ask for. A broad executive intention can explain why a change may be considered; it cannot tell you whether your team has adopted a system, whether it works reliably, or whether a role will be redesigned.

Sources: Work Change Report: AI Is Coming to Work; Skills on the Rise in 2025

What can a task-level comparison tell a worker?

The International Labour Organization’s 2025 index studies potential exposure at the task level, a different question from LinkedIn’s skill and hiring signals. Its task assessments drew on 1,640 employed respondents in Poland, who assessed how generative AI could affect tasks in their own jobs. The researchers used those assessments to inform task-level scoring, extended the analysis through expert review, and applied model scoring across occupations to produce global estimates. The one-in-four and 3.3% figures are therefore modeled global estimates built from task assessments and expert and model judgments—not direct counts from a globally representative worker survey. The ILO estimates that one in four workers globally is in an occupation with some exposure, while 3.3% of global employment falls in its highest exposure category. Its authors judge transformation more likely than full automation because most occupations include tasks that require human input. These estimates indicate potential exposure under the index’s method; they do not measure realized adoption or predict individual redundancy.

That result is not a forecast of realized adoption or a person’s redundancy risk. It is a structured estimate of where task capabilities may intersect with work. LinkedIn gives a view of changing skill labels and labor-market signals; the ILO gives a task-oriented exposure framework. Neither alone establishes what an employer will implement, what demand will do, or how work will be divided after adoption.

Consider an illustrative business-writing bundle. A routine first draft may be a plausible place to test a writing tool. Selecting reliable sources, handling confidential information, checking whether a claim is supported, adapting advice to a client, and approving the final message are separate tasks. Their suitability depends on accuracy, workplace rules, and who remains accountable. List a representative week as repeatable digital work, judgment and verification, and relationship or accountability work. That is more actionable than treating a job title as one indivisible unit.

Sources: Work Change Report: AI Is Coming to Work; Generative AI and Jobs: A Refined Global Index of Occupational Exposure

What is a realistic next move if the report describes my work?

Start with a bounded upgrade tied to work you already understand. If a low-risk, repeatable task is permitted, try an approved tool on that task, verify the result, and record whether total effort and quality improve after review. This tests a workflow in your context; it does not prove that an occupation is being automated or that the capability will be valued elsewhere.

Then compare three paths against your actual constraints. Staying and redesigning work may mean learning basic AI use and verification while keeping domain knowledge. An adjacent move may combine your existing experience with a demonstrated workflow that appears in internal opportunities or local postings. A larger change or formal course, certificate, or degree makes sense only when the destination requires it and its prerequisites, cost, training time, salary implications, location, health, and family demands are workable. The report itself cannot choose among these paths.

The durable learning target is not mastery of one interface that may change. It is the ability to frame a problem, understand the relevant data, evaluate an output, apply domain knowledge, and verify work before it affects someone else. A next-week action is enough to begin: log a representative week, choose one permitted task to test, check its output, then compare a handful of relevant job postings or an internal role path before paying for training. That is an editorial recommendation based on the limits of these measures, not an outcome proven by LinkedIn.

Verdict: the Work Change Report establishes that LinkedIn recorded faster profile skill additions and reports signals of AI-related skill activity and employer interest; it also publishes a substantial but insufficiently reproducible forecast for 2030. It does not establish that 70% of skills in your job will be automated, that every employer has adopted AI, or that you will lose work. The exception is practical: if your own employer is already redesigning a large share of your recurring tasks, act on that local evidence sooner. If you want a task-by-task starting point, the free [AI task checker](/ai-job-risk-checker) offers transparent change-pressure signals, not a displacement probability.

Sources: Work Change Report: AI Is Coming to Work; Generative AI and Jobs: A Refined Global Index of Occupational Exposure

Questions readers ask