Forecast · 2035

Jobs AI will replace by 2035

Jobs where most of today's roles are expected to be gone by 2035, earliest first.

60 jobs · average risk 77%

  1. 01TelemarketerOutbound calling is already largely handled by synthetic voices that never tire and never misread a script.97%
  2. 02Data Entry ClerkDocument understanding models read forms, invoices and handwriting faster and cheaper than any keyboard.96%
  3. 03TranscriptionistSpeech recognition now beats human accuracy on most audio, at a fraction of a cent per minute.95%
  4. 04Customer Support Agent (Tier 1)Most first-line tickets are already resolved without a human touching them.92%
  5. 05Bank TellerBranch transactions keep migrating to apps and self-service terminals.90%
  6. 06ProofreaderGrammar and style checking is a solved problem for everyday text; only high-stakes editorial judgement survives.90%
  7. 07Medical CoderCoding from clinical notes is a direct language-model task.89%
  8. 08Stock PhotographerGenerative imagery covers most generic stock needs instantly and on-brief.88%
  9. 09BookkeeperBank feeds, automatic categorisation and reconciliation remove nearly all manual ledger work.91%
  10. 10Travel AgentPlanning agents can assemble, price and rebook itineraries in seconds.89%
  11. 11Retail CashierCheckout-free stores and self-service terminals are already standard.88%
  12. 12Tax PreparerStandard returns are filled by software with growing confidence.88%
  13. 13Manual Software TesterTest generation, execution and triage are among the fastest-automating tasks in software.87%
  14. 14ReceptionistCheck-in kiosks and voice agents handle arrival and routing.87%
  15. 15Insurance UnderwriterRisk scoring from structured data is exactly what models do best.86%
  16. 16Voice ActorSynthetic voices are convincing and cheap for most commercial needs.86%
  17. 17Legal SecretaryDocument assembly, filing and calendaring are being automated in law firms.86%
  18. 18Radio DJSynthetic voices and automated programming can run a station around the clock.85%
  19. 19ParaplannerPlan generation and compliance paperwork are strongly rule-based.85%
  20. 20HR AdministratorPolicy lookups, onboarding paperwork and scheduling automate well.84%
  21. 21CopywriterVolume copy is essentially free now; distinctive positioning and taste are not.83%
  22. 22TranslatorMachine translation is good enough for most commercial text; humans remain for nuance and liability.84%
  23. 23Quality InspectorVision systems inspect faster and more consistently than human eyes.83%
  24. 24Loan OfficerCredit decisioning is largely algorithmic already.82%
  25. 25ParalegalDocument review, discovery and drafting of routine filings are being automated rapidly.82%
  26. 26IT Support TechnicianSelf-healing endpoints and AI helpdesks absorb most tier-1 tickets.81%
  27. 27IllustratorGeneric illustration is commoditised; a strong recognisable style still commands work.80%
  28. 28Warehouse PickerRobotic picking and goods-to-person systems are deployed at scale.80%
  29. 29Market Research AnalystSurvey design, coding and synthesis are heavily automatable.79%
  30. 30Supply Chain PlannerForecasting and replenishment are textbook optimisation problems.79%
  31. 31RecruiterSourcing and screening are automated; persuasion and closing candidates are not.78%
  32. 32Factory Assembly WorkerRepetitive assembly has been automating for decades; AI extends it to variable tasks.78%
  33. 33Data AnalystQuery writing and dashboard building are increasingly conversational.77%
  34. 34Graphic DesignerLayout and asset production is accelerating dramatically, but design judgement still ships the work.76%
  35. 35Executive AssistantScheduling, inbox triage and travel booking are classic agent tasks.76%
  36. 36Web DeveloperStandard marketing and CRUD sites are generated in minutes.75%
  37. 37Junior Software DeveloperRoutine implementation is heavily assisted; systems thinking and ownership still need people.73%
  38. 38Taxi DriverRobotaxi fleets are scaling city by city.75%
  39. 39Video EditorRough cuts, captions and reformatting are automated; narrative rhythm is still craft.74%
  40. 40AccountantPreparation is automating; sign-off, audit and advisory remain regulated human work.74%
  41. 41Financial AnalystModelling and reporting are accelerating, but capital decisions still carry human accountability.72%
  42. 42AnimatorIn-betweening and rough animation automate; direction and timing remain skilled work.72%
  43. 43Project ManagerStatus collection, reporting and scheduling are highly automatable.71%
  44. 44System AdministratorInfrastructure is code and increasingly self-managing.70%
  45. 45News ReporterRoutine reporting is automated; original sourcing and being physically present are not.70%
  46. 46Security GuardCamera analytics reduce headcount, but physical response still needs people.69%
  47. 47Database AdministratorManaged services and autonomous tuning erode routine DBA work.68%
  48. 48PhotographerGenerated imagery takes the commercial mid-market; real events cannot be faked.68%
  49. 49LibrarianSearch and reference queries are absorbed by AI assistants.67%
  50. 50PharmacistDispensing automates; clinical counselling and prescribing authority do not.64%
  51. 51UX DesignerWireframes and flows are generated easily; research and judgement are not.63%
  52. 52Truck DriverHighway autonomy is arriving on fixed corridors; first and last mile lag behind.66%
  53. 53Real Estate AgentSearch, valuation and paperwork automate; negotiation and local trust persist.65%
  54. 54BaristaAutomated cafés exist, though hospitality and craft keep demand alive.62%
  55. 55Event PlannerLogistics automate; vendor relationships and live problem solving do not.60%
  56. 56Sales Representative (B2B)Prospecting automates; complex, multi-stakeholder deals still close human to human.59%
  57. 57Delivery CourierSidewalk robots and drones cover parts of last mile, but stairs, doors and people remain hard.58%
  58. 58ActuaryStatistical modelling automates, but regulated sign-off and judgement remain.57%
  59. 59RadiologistDetection is superhuman in narrow tasks; accountability, intervention and ambiguity keep humans central.55%
  60. 60Product ManagerDocumentation and analysis compress, but taste, politics and prioritisation remain human.52%

Don't see your job?

Good to know

Jobs AI Will Replace by 2035: Risks and Career Options

This page forecasts jobs whose roles are mostly expected to disappear by 2035, with an average risk of 77% across the 60 jobs listed. These are forecasts, not guaranteed deadlines: exposure depends on the tasks involved, the workplace and whether employers can reliably automate the whole workflow.

What the most exposed jobs have in common

The jobs most exposed to AI tend to process predictable inputs into repeatable outputs. Data entry clerks move information between systems; transcriptionists turn recordings into text; bookkeepers categorize transactions and reconcile accounts. Speech recognition, document extraction and accounting automation can absorb much of that work.

Similar pressure affects telemarketers and tier 1 customer support agents, whose conversations often follow scripts. Copywriters, translators and stock photographers face tools that generate usable first drafts or images. Exposure rises when employers can check outputs cheaply and customers accept standardized results.

Why the same job can face different risks

A job title does not reveal how easily the work can be automated. A travel agent handling straightforward bookings is more exposed than one managing disrupted group travel with complex requirements. A manual software tester repeating scripted checks faces different pressure from someone investigating unpredictable failures.

Regulation, liability, customer trust and messy systems can slow replacement. Medical coders and insurance underwriters may still need to resolve ambiguous records or defend decisions. Retail cashiers and quality inspectors also depend on physical equipment, store layouts and production conditions, so software capability alone does not determine adoption.

How to plan your next career move

Start by separating your work into routine production, exception handling and accountable decisions. Learn the automation tools your employer actually uses, then build evidence that you can catch errors, resolve unusual cases and improve workflows rather than simply produce more output.

Look for adjacent roles that preserve your domain knowledge: bookkeeping can lead toward financial controls, customer support toward complex escalations, and manual testing toward test automation or exploratory testing. Check local vacancies before paying for training. Favor roles with real decision authority, relationships or hands-on responsibilities, while remembering that none is automatically AI-proof.

Questions people ask

Which jobs are most likely to be replaced by AI by 2035?
This page places telemarketers, data entry clerks and transcriptionists among the most exposed roles, with forecast risks of 97%, 96% and 95% respectively by 2027. Their work is especially vulnerable because much of it involves repeatable conversations, structured information or converting speech into text.
Does a high AI replacement risk mean every job will disappear?
No. A forecast of widespread replacement does not mean every employer will eliminate the role on the same schedule; some may retain smaller teams for exceptions, oversight or customers who need human assistance.
Should I change careers if my job is on this list?
Treat the list as a reason to investigate, not a reason to resign immediately. Watch whether your employer is automating your core tasks, examine hiring demand for adjacent roles and build relevant skills while you still have income.
What jobs can't AI replace?
No job is guaranteed to be immune. Work involving unpredictable physical environments, trusted relationships and responsibility for consequential decisions is generally harder to automate end to end, although AI may still change many of its tasks.