Ten jobs that will be much rarer by the end of 2026

People who say AI will destroy all the jobs are wrong. People who say it is not destroying any are also wrong. The FT put a number on it: more than 180,000 corporate job losses have been linked to AI since May 2023, 112,000 of them in 2026 alone, according to the outplacement firm Challenger, Gray and Christmas. US technology companies have cut nearly 140,000 jobs this year while spending record sums on AI. And the World Economic Forum's Future of Jobs Report expects the fastest-declining roles to be clerical and administrative.

So here is a list. Not jobs that will vanish to zero by December, but jobs where there will be clearly fewer people doing them by the end of 2026 than there were at the start of 2024. I have tried to stick to roles where the evidence is already in.

1. Tier-one customer service agents

This is the most visible case. Chatbots now handle the password resets, order tracking and refund requests that used to fill call centres. Sam Altman has said plainly that AI would eliminate some job categories, such as customer service. The twist: Block, after using AI to cut thousands of roles, is running a pilot to rehire humans for customer service. The job that survives is the escalation specialist, not the first-line agent.

2. Junior software developers doing routine coding

Coding assistants have moved from autocomplete to writing whole features. The FT's analysis of whether software engineers survive agentic AI found the profession being reshaped rather than erased, but the entry rung is the one being kicked out. As one FT piece put it, take coding away from a junior developer and you are left with very little. Senior engineers who review, architect and debug are busier than ever.

3. Data-entry and back-office processing clerks

Invoice matching, form processing, claims intake. These were already being offshored; now they are being automated wherever they were offshored to. This is the WEF's fastest-declining category and the least controversial item on this list.

4. Entry-level analysts in professional services

The first two years of a consulting, accounting or banking career used to be spent building slides and models. That work is now largely done by tools, which is why the FT is writing about a graduate 'jobpocalypse' and why PwC's UK graduate applications jumped 35 per cent while places shrank. Consulting firms are even calling juniors back to the office because the human skills, not the technical ones, are what they now need to teach.

5. Paralegals and document reviewers

Contract review and discovery were the first legal tasks to be automated properly. The roles left are supervisory. The FT's Big Question series on office jobs in the age of AI singled out junior professional roles as most exposed.

6. Copywriters for routine marketing

Product descriptions, ad variants, SEO articles. The FT found agencies worried about a real threat to junior creative jobs and about a flattening of creative quality. Senior creatives who set direction are fine. People who produced volume are not.

7. Translators and subtitlers for standard content

Machine translation is now good enough for manuals, support pages and most subtitling. Human translators remain for legal, literary and high-stakes work, but the volume market is gone.

8. Recruitment coordinators and screeners

CV screening, scheduling and first-round filtering are automated. Recruiters who close candidates still exist; the people who booked the interviews mostly do not.

9. Bookkeepers at small firms

Bank-feed categorisation, reconciliation and VAT returns now happen inside accounting software. Accountants who advise are busier; bookkeepers who key in transactions are being replaced by a subscription.

10. Middle managers whose job was reporting

The layer of management that existed to gather status, compile reports and pass them up is being flattened. The FT's coverage of a white-collar recession notes the hiring slowdown is hitting exactly this cohort. Block's decision to cut nearly half its workforce was as much about layers as about tasks.

The honest caveats

Not all of this is AI. The FT is careful to point out that offshoring, post-Covid budget discipline and weak growth explain a lot of the entry-level squeeze. Some economists think the CEO claims are exaggerated. And there is a counter-trend: the FT has argued that AI isn't destroying entry-level jobs so much as changing them, with some firms realising young staff bring AI fluency the older ones lack.

What I am confident of: by the end of 2026 the roles above will employ fewer people, the people who remain will be more senior, and the ladder into those professions will have lost its bottom rungs. That last part is the real problem, and nobody has solved it yet.