AI JOBS & INCOME
AI and Entry-Level Jobs in the UK
The strongest early labour-market concern around generative AI may not be mass redundancy. It may be the shrinking or redesign of the junior tasks through which people traditionally learn a profession.
The clearest UK signal so far
Entry-level hiring is weak in the UK, especially in several information-processing occupations, but the available evidence does not show that AI is the sole cause. A 2026 government and LinkedIn snapshot found 30 of 38 tracked entry-level occupations declining, while also finding that the broader labour market was weak. Accountant, graphic designer and software engineer were among the faster-falling entry-level groups.
For somebody trying to enter work, the practical issue is that AI can change the career ladder before it eliminates a profession. If experienced staff use tools to perform drafting, research, basic coding or routine analysis themselves, employers may need fewer people whose main purpose was previously to do those starter tasks.
Why junior work is structurally exposed
Many professions train people by giving them bounded, lower-risk tasks before they handle harder cases. A junior developer fixes small bugs and writes routine code. A trainee accountant prepares schedules and reconciliations. A junior designer produces variations and production assets. A new analyst gathers information and drafts first-pass material. Those tasks are valuable partly because they create output and partly because they teach the worker how the profession operates.
Generative AI is unusually relevant to that layer because it is strongest at many bounded digital tasks. The employer may not need to believe AI can replace the senior professional. It only needs to believe a senior professional with AI can handle more of the junior workload. That changes hiring economics even if the total volume of work grows.
What Stanford's 2026 research found
Researchers at Stanford's Digital Economy Lab analysed high-frequency payroll data covering millions of US workers. They reported no evidence of widespread economy-wide displacement, but found that employment of workers aged 22 to 25 in the most AI-exposed occupations was substantially weaker relative to less-exposed peers. More experienced workers did not show the same pattern.
A separate Stanford study covering 41 countries examined firms adopting generative AI. It found a decline in the junior share of employment relative to comparable firms, driven mainly by stronger growth in senior employment rather than a collapse in junior headcount. That finding matters because it suggests a possible transition in team composition rather than a simple story of “AI fires everyone”.
Both studies are early evidence. They do not establish that every company will follow the same path, and the US result should not be treated as a UK statistic. But they provide a mechanism worth watching alongside the UK hiring data.
The apprenticeship problem inside white-collar work
If routine starter work disappears, professions face a long-term question: how do people become senior if fewer juniors are hired? An organisation can temporarily solve a productivity problem by giving AI to experienced workers, yet still create a future talent problem if it stops developing new people. That could lead some employers to redesign graduate schemes around supervision, client work and higher-level judgement earlier in a career.
The UK's AI labour-market survey found that many organisations already report AI skills gaps and rely heavily on on-the-job learning. That makes practical experience more valuable, not less. The challenge is gaining that experience when traditional starter tasks are automated or expected to be completed much faster.
How to compete for entry-level work when AI is normal
Employers need evidence that you can define a problem, check sources, notice errors and explain decisions — not merely produce polished text or code that a tool could generate.
Where appropriate, learn the tools used in the field and be ready to explain how you verified the result, protected confidential information and knew when not to use them.
Generic prompting is easy to copy. Understanding the industry, customer, regulation or technical system gives the AI output context and makes mistakes easier to spot.
Placements, apprenticeships, volunteering, personal projects and part-time work can demonstrate reliability and judgement when a CV otherwise looks identical to hundreds of applicants.
Do not assume every digital career is closing
The same technology can reduce some tasks and increase demand elsewhere. The UK AI labour-market survey reports recruitment difficulties for experienced AI talent and growing demand for people who can deploy and understand the technology. Skills England's 2026 digital assessment describes demand shifting away from some routine coding and testing toward oversight, assurance, judgement, communication and effective use of AI alongside technical expertise.
That means a student choosing computing, design, finance or another exposed field is not automatically making a mistake. The stronger strategy is to understand how the role is changing and avoid building an identity around the easiest-to-automate layer of the work.
If the bottom rung disappeared too quickly
Imagine firms across several professional sectors deciding that experienced workers with agents can perform most junior production work. Graduate vacancies fall sharply for three recruitment cycles. Applicants respond by collecting more qualifications, but employers still ask for practical experience. Competition intensifies for internships and apprenticeships, and a generation takes longer to enter stable professional work.
The system eventually adapts with redesigned training roles, but the people entering during the transition carry the cost. This scenario is not a forecast; it illustrates why the shape of junior hiring matters even if total employment remains relatively stable.
A practical plan for graduates and career starters
Track real vacancies for the job you want rather than relying on broad career rankings. Save a sample every few months and compare the tasks, software and experience requirements. If the role changes, your preparation should change with it. Speak to people doing the work and ask what junior staff now spend time on compared with two years ago.
Keep applications focused. A hundred generic AI-written applications can be weaker than ten that demonstrate actual knowledge of the employer and role. If using AI to help draft, rewrite the result until it reflects real experience and can be defended in an interview. Never invent projects, skills or results.
Finally, keep the household side realistic. Entry into work can take longer in a weak market even without AI. Know your monthly costs, avoid expensive courses with vague placement claims and consider interim work that preserves income while you continue building relevant evidence.