
Lists of “jobs AI will replace” are everywhere, and most of them quote percentages nobody can trace. This one is built only on studies you can check yourself: the World Economic Forum’s Future of Jobs Report, the Tufts American AI Jobs Risk Index, Stanford’s AI Index 2026, Brookings’ research on workers’ ability to adapt, and Microsoft’s analysis of 200,000 real Copilot conversations. Where the data is weak, I say so.
Quick Answer: Which Jobs Are Most at Risk From AI?
The jobs most at risk from AI by 2027 are routine office and information roles: data entry clerks, customer service representatives, administrative assistants, bank tellers and bookkeeping clerks. Next come writers, translators and HR assistants, whose core tasks AI already performs well. Junior software developers and graphic and web designers are the fastest-rising risk. Physical and care work remains the safest: see our list of the 10 jobs AI can’t replace.
| # | Job | Risk level | What AI takes over | Key evidence |
|---|---|---|---|---|
| 1 | Data entry clerks | Very high | Typing, form filling, document extraction | WEF: among the fastest-declining jobs |
| 2 | Customer service representatives | Very high | Routine questions, chat and phone tickets | Microsoft top 10; Stanford: early job decline |
| 3 | Administrative assistants & secretaries | Very high | Scheduling, email, minutes, documents | WEF; Brookings: low ability to adapt |
| 4 | Bank tellers & related clerks | Very high | Transactions, account questions | WEF: near the top of the declining list |
| 5 | Bookkeeping, accounting & payroll clerks | Very high | Invoice coding, reconciliation, payroll | WEF; Brookings: payroll clerks |
| 6 | Writers & copywriters | High | Drafts, product copy, rewrites | Tufts: 57% displacement |
| 7 | Translators & interpreters | High | Written translation, first drafts | Microsoft: #1 most AI-applicable job |
| 8 | HR assistants & recruiting coordinators | High | CV screening, scheduling, onboarding admin | ILO, Brookings, PwC |
| 9 | Junior software developers | Rising | Boilerplate code, tests, simple features | Stanford: −20% for ages 22–25; Tufts: 55% |
| 10 | Graphic & web designers | Rising | Layouts, banners, variations, mockups | WEF: new on declining list; Tufts: 55% |
How This List Was Built
There is no single “AI risk score” that every researcher agrees on, and the studies measure different things:
| Source | What it measures | Scope |
|---|---|---|
| WEF Future of Jobs 2025 | Which jobs employers expect to shrink by 2030, for all reasons | 1,000+ employers, 55 economies |
| Tufts AI Jobs Risk Index | How likely AI exposure turns into job loss in 2–5 years | 784 U.S. occupations, March 2026 |
| Stanford AI Index 2026 | Actual employment changes by age group | U.S. payroll data, April 2026 |
| Brookings / GovAI | AI exposure combined with ability to find a new job | U.S. workers, January 2026 |
| Microsoft “Working with AI” | How much of a job’s tasks people already do with AI | 200,000 Copilot conversations, July 2025 |
Because these scales cannot be honestly merged into one number, the jobs are grouped into three risk levels instead of a fake precise ranking. Very high means almost every source points at the job. High means AI already performs the core task well. Rising means the job only recently appeared on the lists, and the damage is concentrated on entry-level roles.
Two jobs high on the WEF list are deliberately left out: postal service clerks and retail cashiers. They are declining fast, but mainly because of digital services and self-checkout, not AI.
Very High Risk: Routine Office and Information Work
1. Data Entry Clerks
Data entry is the clearest case. The work is copying information from one place to another, and AI can now read invoices, forms, emails and scanned documents and put the result straight into a system. The WEF lists data entry clerks among the fastest-declining jobs through 2030.
What is left for humans: checking exceptions and data quality. That is a smaller team, and it usually needs a different title, such as data operations or data quality analyst.
2. Customer Service Representatives
Customer service representatives rank sixth in Microsoft’s list of occupations where AI applies most to the work people do. Stanford’s AI Index 2026 reports that the drop in employment among young workers in AI-exposed jobs “repeats in other jobs with higher levels of AI exposure, like customer service.” Chatbots and voice agents now handle password resets, order tracking and refunds.
What is left for humans: complicated complaints, angry customers, sales and retention. Agents who can handle escalations are safer than those who answer scripted questions.
3. Administrative Assistants and Executive Secretaries
Calendars, email triage, meeting notes and document drafts are exactly what AI assistants are built for, and the WEF lists administrative assistants and executive secretaries among the fastest-declining jobs. This group also faces a second problem. Brookings found that about 6.1 million U.S. workers combine high AI exposure with low capacity to adapt, concentrated in clerical and administrative roles, and about 86% of them are women. Secretaries and administrative assistants alone account for 1.7 million of those workers.
What is left for humans: coordinating people, handling confidential situations and acting as an executive’s right hand. Assistants who move toward office management or operations have the best options.
4. Bank Tellers and Related Clerks
Bank tellers and related clerks are among the very first jobs on the WEF’s list of the fastest-declining roles. Mobile banking started this decline long before generative AI, and AI assistants inside banking apps now answer the account questions that used to bring customers to a branch.
What is left for humans: advisory and sales roles in branches, plus fraud and vulnerable-customer cases.
5. Bookkeeping, Accounting and Payroll Clerks
Coding invoices, reconciling accounts and running payroll are rule-based tasks, and accounting software now does more of them automatically. The WEF lists accounting, bookkeeping and payroll clerks among the fastest-declining jobs. Brookings names payroll and timekeeping clerks among the most exposed workers with the lowest ability to adapt.
What is left for humans: judgment on tax, cash flow and unusual transactions. A bookkeeper who becomes an advisor to small businesses is far safer than one who only enters numbers.
High Risk: AI Already Does the Core Task
6. Writers and Copywriters
In the Tufts index, writers and authors have the highest displacement rate of any occupation: 57%. Writers and authors also rank fifth in Microsoft’s list, with editors, proofreaders and technical writers not far behind. Product descriptions, routine blog posts and ad variations are the first to go.
What is left for humans: original reporting, expertise, a recognizable voice and editing AI output. In our own field, Google rewards content with real experience behind it, which is why we wrote about how to write SEO content that ranks and gets cited.
7. Translators and Interpreters
Interpreters and translators are number one in Microsoft’s ranking of the occupations where AI applies most to the actual work. Machine translation now produces usable first drafts in seconds, and many translators already work as post-editors of AI output.
What is left for humans: certified legal and medical translation, live interpreting in sensitive settings, literary work and localization that needs cultural judgment.
8. HR Assistants and Recruiting Coordinators
HR splits into two very different groups. Transactional work such as screening CVs, scheduling interviews, onboarding paperwork and payroll is highly exposed. An ILO and NASK study cited by HRM put payroll clerks in the highest exposure category, while placing HR managers in the minimal exposure group. PwC’s 2026 AI Jobs Barometer, as reported by Recruiting News Network, describes recruiters shifting from screening toward negotiation and relationship work.
One caution: there is no study yet measuring actual HR layoffs caused by AI. Most figures come from employer surveys about plans, so treat them as forecasts.
What is left for humans: hiring decisions, employee relations, conflict and change management.
Rising Risk: Newer to the Lists, Hitting Juniors First
9. Junior Software Developers
This is the most measured change so far. Stanford’s AI Index 2026 reports that employment among software developers aged 22 to 25 has fallen nearly 20% since 2024, “even as their older colleagues’ headcount grows.” Tufts gives computer programmers a 55% displacement rate. Companies are hiring fewer juniors because AI coding tools do much of the work juniors used to learn on.
What is left for humans: system design, code review, security and understanding what the business actually needs. We looked at this in depth in Will AI Replace Web Developers and App Creators? and in what goes wrong when people vibe code a business website.
10. Graphic and Web Designers
In its 2025 report the WEF, for the first time, placed graphic designers and legal secretaries among the fastest-declining jobs, a change it links to generative AI. Tufts gives web and digital interface designers a 55% displacement rate. Banners, social graphics, layout variations and simple mockups are now generated in seconds.
What is left for humans: brand strategy, UX research, art direction and design systems. Designers who own the “why” are much safer than those who produce variations.
Jobs Just Outside the Top 10
These roles also appear high in at least one major study and are worth watching:
- Paralegals and legal secretaries: named by the WEF as newly declining.
- Telemarketers and telephone operators: high in Microsoft’s list, though the jobs were shrinking long before AI.
- Proofreaders, editors and technical writers: all in Microsoft’s top 25.
- Market research analysts: named by Tufts among the largest income losses and listed by Microsoft.
- News reporters and journalists: in Microsoft’s top 20.
What the Data Does Not Show
“At risk” does not mean “gone by 2027.” Three caveats matter:
- Exposure is not job loss. Microsoft’s researchers write that their data “does not indicate it can fully perform any single occupation.” Brookings calls its findings directional, not predictive.
- The damage so far is mostly fewer hires. The clearest measured effect is that young people find it harder to get a first job in exposed fields, not that experienced staff are fired en masse.
- AI is not the main reason for layoffs. Challenger, Gray & Christmas counted 121,401 announced U.S. job cuts citing AI through October 2026, about 20% of all cuts. Market and economic conditions have overtaken AI as the leading reason, with 130,531 cuts. Companies also cite AI when cutting for other reasons.
The big picture is also not purely negative. The WEF expects 170 million jobs to be created and 92 million displaced by 2030, a net gain of 78 million, though the new jobs are not in the same places or for the same people.
How to Protect Your Career if Your Job Is on This List
- Learn the AI tools of your own job first. The person who uses AI to do the work of three is usually the one who stays.
- Move toward the parts AI handles badly: exceptions, judgment, negotiation, accountability and client relationships.
- Get closer to the decision. A bookkeeper becomes a financial advisor, a copywriter a content strategist, a recruiting coordinator a recruiter.
- Build proof of expertise in your name: a portfolio, certifications and published work. AI copies average output, not reputation.
- Consider a move to a safer field. Healthcare and the licensed trades lead our list of jobs AI can’t replace.
- Keep a financial buffer. Brookings found that savings, age, local opportunities and skills breadth decide how hard a job loss hits.
Frequently Asked Questions
Which jobs will AI replace first?
Routine office and information jobs: data entry clerks, customer service representatives, administrative assistants, bank tellers and bookkeeping clerks. Every major dataset points at this group, and the clearest early damage is fewer openings for junior staff rather than mass firings.
How many jobs will AI replace by 2027?
No one can give an exact number. Tufts estimates about 9.3 million U.S. jobs at risk of displacement within two to five years, in a range of 2.7 to 19.5 million. Globally, the World Economic Forum expects 92 million jobs displaced and 170 million created by 2030, from all causes, not only AI.
Will AI replace HR jobs?
Not all of them. Transactional HR work such as CV screening, interview scheduling, onboarding paperwork and payroll is highly exposed. HR managers and HR business partners, whose work is judgment, negotiation and people, are in the low-exposure group.
Will AI replace programmers?
It is hitting entry-level developers first. Stanford reports employment of software developers aged 22 to 25 fell nearly 20% since 2024, while older colleagues kept growing. Experienced developers who design systems and review AI output are still in demand.
What jobs are safest from AI?
Physical, hands-on and care work. Tufts classes 38% of American workers as AI-proof, including roofers, orderlies and dishwashers, though many of those jobs are low paid. Trades, nursing and personal care rank among the least exposed. See the full list of jobs AI can’t replace.
Is AI the main reason for layoffs in 2026?
No longer. Challenger, Gray & Christmas counted 121,401 announced U.S. job cuts citing AI through October 2026, about 20% of all cuts, but market and economic conditions overtook AI as the top stated reason.
Sources
- World Economic Forum, Future of Jobs Report 2025 (press release, January 2025)
- World Economic Forum, Future of Jobs Report 2025: Jobs outlook
- Tufts Digital Planet, American AI Jobs Risk Index (March 24, 2026)
- Stanford HAI, Inside the AI Index: 12 Takeaways from the 2026 Report (April 13, 2026)
- Brookings, Measuring US workers’ capacity to adapt to AI-driven job displacement (January 21, 2026)
- Tomlinson et al., Working with AI: Measuring the Occupational Implications of Generative AI (Microsoft Research, 2025)
- Challenger, Gray & Christmas, October 2026 Job Cuts Report (November 5, 2026)
- HRM, Are HR managers facing job loss with generative AI? (ILO and NASK study)
- Recruiting News Network, How AI’s two-track divide is going to disrupt HR jobs



