Will AI take your job? Here's what's really changing, and what to do about it.
Last updated: October 2026
AI is not wiping out whole jobs overnight. It is taking over specific tasks, mostly routine screen-based ones, and employers are hiring fewer juniors for roles built around those tasks. This programme shows you, role by role, which of your tasks are changing, which skills keep you valuable, and which growing roles your experience transfers to.
The evidence (As of October 2026)
In the most AI-exposed jobs, employment of 22–25-year-olds is about 19% below where it would otherwise be. The biggest drops are where AI automates tasks rather than assists with them. Source: Stanford Digital Economy Lab, August 2026 revision.
UK job vacancies in automation-exposed roles such as software and consulting fell 37% from late 2022, against 26% for other roles. Source: Bloomberg analysis of ONS data, 2026.
43% of UK employers say AI or automation investment has reduced entry-level roles (60% of large employers). Source: Work Foundation, August 2026.
Hands-on and trade work has held steady. Source: Stanford Digital Economy Lab, 2026.
Not every job cut is caused by AI. Weak demand and costs matter too, and some firms that cut staff for AI have since rehired.
The three moves
Adapt: hand the routine tasks to AI and become the person who checks, improves and owns the result.
Upskill: add the skills AI makes more valuable: judgement, relationships, explaining things, fixing exceptions, and knowing the tools.
Move: shift into a nearby role that is growing, using the experience you already have.
Roles covered
Customer service and call centre advisers: Chatbots and AI agents now answer most simple, repeat questions such as order status, password resets and opening hours. What is left for people is the harder work: angry or vulnerable customers, complaints, exceptions and anything needing judgement. Advisers who can supervise AI, handle escalations and improve the bot's answers are the ones teams keep.
Admin, data entry and office support: Typing data from forms, emails and invoices into systems is one of the tasks AI does fastest and most cheaply. Office roles are not disappearing, but they are shifting from entering information to owning the process: checking the AI's work, fixing exceptions and keeping systems tidy. People who can set up simple automations become the most useful person in the office.
Bookkeeping, payroll and accounts clerks: Bank feeds, receipt scanning and AI matching in software like Xero, QuickBooks and Sage now do much of the routine bookkeeping. The value is moving to reviewing, advising and spotting problems. Bookkeepers who can explain the numbers to a business owner, and keep the system set up correctly, are in a stronger position than ever.
Bank, insurance and back-office operations: Large banks and insurers are among the most open about using AI to cut back-office work such as checks, claims handling and document processing. Standard Chartered has said it plans to cut around 15% of back-office roles by 2030. Staff who understand both the process and the controls around AI are needed to run, check and govern these systems.
Junior software developers: AI coding assistants now write much of the routine code, so firms are hiring fewer pure juniors. Developers are still needed, but the job is moving towards reviewing AI code, understanding the business problem, testing, security and joining systems together. Juniors who can show they ship working, secure software with AI tools stand out.
Junior analysts and consultants: Research, slide-building and first-draft analysis, the traditional junior workload in consulting and professional services, are now quick with AI. Firms are hiring fewer graduates for that work. The value is moving to framing the right question, checking the AI's analysis, and turning findings into decisions clients act on.
Marketing, copywriting and graphic design: AI now produces first drafts of copy, images, social posts and ad variations in seconds, so demand for routine production work is falling. Brand judgement, strategy, taste, original ideas and the ability to measure what works are worth more. Creatives who direct AI, rather than compete with it on volume, are winning the work.
Legal support (paralegals and legal secretaries): AI tools now handle document review, first-draft letters, research summaries and bundle preparation much faster than before. Firms need fewer people for volume document work, but more people who can check AI output for accuracy, manage matters and handle clients. Errors in AI legal research, including invented cases, mean careful human verification remains essential.
Unlikely for complex, emotional or regulated contact. Several firms that cut too far have brought humans back for escalations. Routine contact volumes are falling, though.
Is data entry still a job?
Fewer pure data-entry roles are being advertised; the work is shifting to checking and exceptions.
Will AI replace bookkeepers?
It is replacing much of the data processing. Review, advice and compliance still need people.
Is my role at risk if I'm in operations?
High-volume, rules-based steps are the most exposed. Control, exceptions and customer-outcome roles are less exposed.
Should I still learn to code?
Yes, but aim to be the person who can judge and fix AI code, not just type it.
Is consulting still a good graduate career?
The entry route is narrower; sector expertise and AI fluency matter more.
Is graphic design dying?
Routine production is shrinking; brand, concept and art-direction work is not.
Will AI replace paralegals?
It reduces volume document work. Verification, matter management and client work remain.