The loudest question about artificial intelligence at work is usually the wrong one: “Which jobs will AI eliminate?” The more useful question is which parts of today’s jobs are already being reorganized around AI.
That distinction matters. The International Labour Organization’s latest evidence review says measurable productivity gains from generative AI are real but uneven, while large-scale job displacement remains limited so far. Its occupational exposure index also finds that roughly one in four workers worldwide is in a job with some exposure to generative AI, with transformation more likely than wholesale replacement.
Here are 10 occupations where that transformation is moving especially fast.
1. Administrative assistants
Scheduling, meeting notes, routine correspondence, travel planning and document formatting are increasingly handled with AI copilots. That does not remove the need for assistants, but it changes the premium: judgment, coordination, discretion and knowing which requests actually matter become more valuable than typing or formatting speed.
2. Bookkeepers and accounting clerks
Clerical and bookkeeping work remains among the most exposed categories in ILO research. AI can classify transactions, draft reconciliations, flag anomalies and prepare first-pass reports. Human review still matters because tax treatment, unusual transactions and accountability cannot be delegated to a model without controls.
3. Customer-service representatives
AI systems can now summarize long customer histories, propose replies and answer routine questions before a human agent enters the conversation. The job is shifting toward escalations, emotionally difficult cases and situations where policy has to be interpreted rather than simply quoted.
4. Software developers
Programming is one of the most visible examples of AI augmentation rather than simple automation. Developers increasingly use models to generate boilerplate, explain unfamiliar code, write tests and troubleshoot. The bottleneck moves toward architecture, security, product decisions and verifying whether generated code is actually correct.
5. Financial analysts
ILO researchers have noted rising exposure in highly digitized professional roles, including financial analysts and investment advisers. Models can summarize filings, compare scenarios and extract signals from large document sets. But the expensive mistakes in finance usually come from bad assumptions, weak data or misplaced confidence — all areas where human accountability remains central.
6. Marketing and advertising specialists
Drafting ten versions of copy, creating campaign variants and summarizing audience research can now take minutes. The harder part is deciding what a brand should say, what evidence supports the claim and what will damage trust. AI reduces production friction; it does not automatically create good positioning.
7. Journalists, editors and media producers
Generative AI can transcribe interviews, summarize documents, suggest headlines and create rough visual or audio assets. ILO’s 2025 update found that advances in image, voice and video generation increased exposure scores for media and web-related occupations. Verification, original reporting, source judgment and editorial responsibility become more—not less—important when synthetic content gets cheaper.
8. Lawyers and paralegals
Document review, contract comparison and first-pass legal research are natural targets for language models because so much legal work is text-heavy. The change is most obvious in the early stages of a task: locating clauses, organizing precedents and drafting. Legal strategy, client advice and responsibility for errors remain human obligations.
9. Recruiters and HR specialists
AI can draft job descriptions, summarize résumés, structure interview notes and surface patterns across large applicant pools. That also raises risks around bias, privacy and opaque screening. The profession may spend less time moving information between systems and more time validating decisions and explaining them.
10. Teachers and trainers
Lesson-plan drafts, quizzes, differentiated exercises and feedback templates are increasingly easy to generate. The role of the teacher is harder to automate: diagnosing why a student is stuck, motivating a class, setting standards and deciding whether an answer demonstrates real understanding. The technology changes preparation time faster than it changes the human relationship at the center of learning.
The pattern: tasks disappear before occupations do
OECD research published in 2026 describes three channels through which AI changes labor markets: automating existing tasks, creating new tasks and occupations, and raising productivity. High-skill jobs can be highly exposed to AI while still being difficult to automate because they depend on non-routine judgment and social skills.
That helps explain why “AI exposure” should not be read as “job loss.” Exposure means a meaningful share of a job’s tasks can be affected. Whether employment falls depends on what employers do with the saved time, whether demand grows and whether workers can move into the higher-value parts of the role.
What workers should watch now
The practical dividing line is becoming less about whether someone uses AI and more about whether they can supervise it. Workers who can check outputs, provide context, protect sensitive information and connect model results to real-world decisions are likely to have an advantage over those who treat AI as either magic or a threat to avoid.
The World Economic Forum’s Future of Jobs Report estimated that broad labor-market shifts—not AI alone—could create 170 million jobs and displace 92 million by 2030. The numbers are forecasts, not a timetable. The nearer-term evidence is more modest: jobs are being rewritten task by task.
Sources: International Labour Organization, The impact of GenAI on jobs, productivity and work organization (June 2026); ILO-NASK, Generative AI and Jobs: A Refined Global Index of Occupational Exposure (2025); OECD, Skills in the AI age (July 2026); World Economic Forum, Future of Jobs Report 2025.
