INTERACTIVE TOOL // LABOUR-SHOCK STRESS TEST
AI Job Exposure Checker
The question is not only “will AI replace my job?” Jobs are bundles of tasks. Score the work you actually do and see which parts could face the strongest AI pressure — then stress-test what that could mean for household income in a faster, uglier transition.
SCENARIO: autonomous agents become cheap enough to absorb large volumes of digital work within months. Employers freeze recruitment, compress teams and automate routine output faster than retraining systems can react. Your occupation still exists — but the market value of half its tasks changes before your household budget does.
This checker estimates task exposure, not redundancy probability. Exposure can mean automation, augmentation or changed skill requirements.
Describe the task mix
Choose how much of a normal working week each category represents. The options are deliberately broad; this is a stress test, not an occupational classification.
Your AI task-exposure result
Score the eight task categories to estimate where AI pressure may be strongest.
Exposure is not the same thing as losing a job
International research on generative AI exposure repeatedly makes this distinction. An occupation may contain many tasks that AI can assist with while still requiring human judgement, responsibility, physical presence or interpersonal work. The result may be faster output, fewer junior tasks, changed staffing ratios, new roles or, in some cases, displacement. A task-level checker cannot tell you which outcome an employer or economy will choose.
The ILO's 2025 refined global index estimates that roughly one in four workers worldwide are in occupations with some degree of generative-AI exposure, while only a much smaller share fall into the highest-exposure category. Its central conclusion is that transformation is more likely than complete replacement for most jobs because human input remains necessary. That is why this tool reports exposure rather than pretending to know your redundancy odds.
Where the pressure could arrive first
Highly digitised, repeatable information tasks are easier for current generative systems to touch than work requiring physical manipulation in messy environments. Drafting standard text, routine administration, basic research, coding assistance and digital production can all experience significant AI pressure. Physical repair, direct care, real-world inspection, accountable decisions and complex human negotiation contain different bottlenecks.
But low exposure today is not a permanent shield. Better robotics, stronger multimodal models and more capable agents could move the boundary. The useful household question is therefore not “am I safe forever?” It is “which parts of my income depend on tasks that could become cheaper or faster to automate, and what would I do if demand for those tasks changed quickly?”
The fictional doomsday labour scenario
Take the transition to its worst plausible fictional extreme. Agent systems become capable enough to run large portions of office workflows with minimal supervision. Companies cut layers of routine work simultaneously. Freelance markets flood with machine-generated output. Entry-level routes into professional careers narrow. At the same time, misinformation and automated trading shocks destabilise confidence. Households do not experience “AI” as a robot at the door; they experience it as a payslip that stops arriving.
That is why income resilience belongs on a survival site. Emergency cash, lower fixed commitments, transferable skills, copies of qualifications, offline records and the ability to keep operating during digital disruption can all buy time. The point is not to panic about a percentage. It is to identify concentration risk before a fast economic transition makes the household discover it involuntarily.
Treat exposure as task-level, not job-level
Most jobs contain a mixture of tasks. Some may be highly automatable while others depend on physical work, trust, judgement, relationships or local knowledge. Use the checker to identify which parts of a role could change rather than assuming the entire occupation either disappears or remains untouched.
This makes the result more useful for planning skills and career moves.
Look for complementary skills
If AI can perform part of the work faster, the opportunity may be to become better at supervising, verifying or combining that output with domain knowledge. Communication, practical judgement, accountability and the ability to work with imperfect real-world information can remain valuable even as software takes over routine tasks.
The tool should prompt questions about adaptation rather than predict a personal employment outcome.
Recheck as tools change
AI capability and workplace adoption can move quickly, so a score is a snapshot. Re-run the checker when the tools used in your industry change or when your role shifts toward different tasks.
Use the result alongside real job-market evidence, employer behaviour and your own experience rather than treating it as a forecast.
Build an income fallback before it is needed
Task exposure matters most when a household depends heavily on one employer, one type of client or one narrow set of digital tasks. A sensible response is to strengthen options while income is still stable: keep an up-to-date record of qualifications and work samples, know which skills transfer into adjacent roles, and understand the minimum monthly costs the household would need to cover during a disruption.
That is different from assuming an occupation is doomed. Labour markets adapt unevenly, and adoption can be slower or faster than headlines suggest. The point of a stress test is to make a sudden change less damaging if it arrives, not to treat a model score as a forecast of personal redundancy.
Continue from here
Build this topic into a wider plan, then see how the same dependency could fail in an AI-collapse scenario.
Evidence baseline
The ILO assesses exposure at task and occupation level and stresses that exposure does not automatically mean replacement; transformation is the more likely outcome for most occupations in its current analysis.