AI JOBS & INCOME
Will AI Take My Job?
No honest tool can tell one person that AI will or will not take their job. What can be assessed is the task mix inside the role, how quickly employers are adopting AI, whether hiring is changing, and how much room the household has to absorb an income shock.
The immediate answer
AI is already changing tasks and hiring, but current evidence does not support a simple claim that every exposed job will disappear. The International Labour Organization estimates that about one in four workers globally are in occupations with some degree of generative-AI exposure, while stressing that transformation is more likely than complete replacement for most jobs because human input remains necessary.
The useful question is therefore not “can an AI do something I do?” It is “how much of my paid work is made of tasks that are becoming cheap to automate, how valuable are the remaining human tasks, and how quickly is my employer or industry changing?”
Start with tasks, not job titles
Two people with the same job title can have very different exposure. One accountant may spend much of the week preparing routine reports, reconciling records and drafting standard explanations; another may spend most of the week dealing with unusual cases, client trust, regulated judgement and responsibility for decisions. A language model may touch both jobs, but the likely effect on their work can be very different.
The ILO's refined exposure index uses thousands of occupational tasks rather than treating an occupation as one indivisible block. That is the right mental model for personal planning too. List what you actually do in a normal week. Mark the tasks that are mainly digital, repeatable and based on information already available in a computer. Then mark the tasks that depend on physical presence, responsibility, relationships, local knowledge, negotiation or dealing with messy exceptions.
Run the site's AI Job Exposure Checker after doing this exercise. Treat the score as a prompt for questions, not a redundancy probability.
What the UK evidence says so far
UK data shows a labour market under pressure, but it does not justify blaming every weak hiring number on AI. A June 2026 government and LinkedIn snapshot found that entry-level hiring was weak and that 30 of 38 tracked entry-level occupations were declining. Information-processing roles such as accountants, graphic designers and software engineers were among the fastest-declining groups. The same report explicitly cautioned that entry-level hiring was falling broadly in line with the wider labour market and that more research was needed before attributing the pattern to AI.
That distinction matters. Interest rates, business confidence, sector cycles, post-pandemic adjustments and ordinary cost cutting can all affect hiring. AI can interact with those forces without being the sole cause. A worker who assumes every vacancy decline is proof of imminent automation may make a poor decision for the wrong reason.
At the same time, there are enough signals to justify preparation. Stanford researchers using US payroll data through June 2026 found no evidence of widespread economy-wide displacement, but did find a substantial relative employment gap for workers aged 22 to 25 in highly AI-exposed occupations. A separate 41-country study found that firms adopting generative AI reduced the junior share of their workforce relative to comparable firms, largely because senior employment grew faster. Neither result means “AI has eliminated junior work”, but both make entry-level pathways worth watching.
Five signs your role deserves a closer look
Text, code, slides, routine analysis and standardised documents are easier for current systems to touch than work that requires physical action in an unpredictable environment.
If tasks follow the same inputs, rules and output format again and again, employers have a clearer target for workflow automation.
Watch what happens to the work once given to trainees or assistants. If senior staff use AI to absorb it, the role may change before the occupation disappears.
Automation moves faster when customers tolerate or prefer self-service, chat interfaces and automated decisions.
The strongest signal is not a headline. It is a real change in team size, hiring, targets, software, job descriptions or the ratio of junior to senior workers.
What makes a role harder to automate completely?
Physical dexterity in unstructured environments remains a bottleneck for many jobs. So do accountability, high-stakes judgement, relationship building, persuasion, care, trust and the ability to handle novel situations when the available information is incomplete. Regulation can also require a human or named professional to remain responsible even when AI assists with part of the process.
None of those qualities makes a career permanently “AI-proof”. Robotics can improve, regulations can change and customers can alter what they accept. But they can make the transition look more like job redesign than job deletion. That is why the most useful career strategy is often to move toward the parts of the work where human value remains complementary to the technology.
If adoption moved much faster than institutions could adapt
Imagine capable agents becoming cheap enough to absorb a large share of routine office work within a year. Companies pause junior recruitment first, then combine roles as experienced staff supervise automated workflows. Freelancers face a sudden flood of low-cost machine output. Retraining programmes expand, but not quickly enough for every worker. The danger to a household is not a robot arriving at the door. It is several months of weaker income while rent, food, debt and childcare costs continue normally.
Scenario: this is a stress test, not a forecast. Current evidence does not show economy-wide mass displacement on this timetable.
Build an income-resilience plan without panicking
Start with the household numbers. Know the minimum monthly cost of keeping housing, utilities, food, transport, insurance and essential commitments running. Separate that figure from discretionary spending. If income stopped or fell sharply, knowing the real minimum turns an abstract employment fear into a measurable runway.
Then strengthen options while still employed. Keep a current CV, copies of qualifications and examples of work that demonstrate judgement rather than only routine output. Identify adjacent roles that use the same domain knowledge. Learn the AI tools entering the industry instead of waiting for a redundancy notice to discover them. If the job is becoming more supervisory, verification-heavy or client-facing, build those skills deliberately.
Finally, reduce single-income concentration where practical. That does not mean starting a dubious “side hustle” because AI headlines are frightening. It may mean maintaining professional contacts, keeping a small emergency reserve, understanding benefits or redundancy rights, and making sure one lost login or employer device does not also remove access to personal records.
What not to do
Do not resign because an online calculator gives a high exposure score. Do not retrain for a supposedly “AI-proof” career without checking demand, entry requirements and local pay. Do not assume an employer using AI is planning layoffs; it may be using the tools to increase output, change service levels or fill skills gaps. And do not confuse a viral forecast about millions of jobs with evidence about your occupation, region and level of experience.
The best response to uncertainty is optionality: stronger skills, clearer household finances, better evidence and more routes into paid work.