AI job impact: frequently asked questions

It depends on the occupation and the tasks within it. Per the ILO 2025 refined Generative AI Occupational Exposure Index, occupations are placed in one of four exposure gradients (Low, Moderate, High, Very High). Per Brookings 2024, task-level disruption is more granular than occupation-level. The honest answer: parts of every knowledge job are exposed; whole-job displacement is rarer than headlines suggest. Run your job through the calculator on the homepage for an occupation-specific gradient, task breakdown, and BLS 2024-2034 projection.

The ILO 2025 refined index places ISCO-08 occupations into four exposure gradients. The highest-exposure gradient includes data-entry clerks, telephone operators, customer service representatives, bookkeeping clerks, and administrative assistants. The list is a measure of task-level exposure to generative AI, not a prediction of which occupations will disappear. See the highest-exposure list at /jobs-ai-will-replace/.

AI agents are a 2024-2026 subset of AI deployment, where LLM-based systems take actions across tools rather than just answering questions. Per LinkedIn data published April 2026, hiring is down approximately 20% since 2022 but LinkedIn states it has not seen AI as the demonstrable cause. The honest answer: agents are extending what LLMs do; the disruption observed so far is task-level, not whole-job. For a primer on AI agents, see whatisanaiagent.com.

The calculator returns the ILO 2025 exposure gradient for the matched occupation, the task-level breakdown per Brookings 2024, and the BLS 2024-2034 employment projection. Risk is a spectrum, not a binary, and the gradient is occupation-family-level, not personalised. Use the per-occupation deep dive at /occupations/{slug}/ for full sourcing.

Per the ILO 2025 refined index, the lowest-exposure gradient covers occupations heavy in physical interaction, social-emotional skills, or judgement-under-uncertainty. Examples include registered nurses, secondary school teachers, electricians, social workers, and skilled trades. Safe is calibrated, not absolute. See /ai-proof-jobs/ for the published lowest-exposure list.

Aggregate exposure estimates: Goldman Sachs 2023 (300 million full-time jobs exposed globally), McKinsey 2024 (30% of current hours worked could be automated by 2030 in the midpoint scenario), Brookings 2024 (over 30% of workers could see at least 50% of their tasks disrupted). These are exposure estimates, not displacement predictions. As of mid-2025, Brookings finds aggregate-labour-market data does not show AI-driven mass unemployment.

Aggregate predictions about displacement have a poor track record. Frey-Osborne 2013 estimated 47% of US employment was at high risk of computerisation; the 2024 unemployment rate did not reflect that projection. Task-level methodology (Brookings 2024) is more defensible than occupation-level prediction. The calculator on this site outputs exposure, not displacement. The full pre-empt is at /how-to-argue-with-this/.

Per WEF Future of Jobs Report 2025, the top growing skills are AI and big data, networks and cybersecurity, technological literacy, creative thinking, and resilience and flexibility. Per BLS Occupational Outlook 2024-2034, the projected-fastest-growing occupations include AI and machine-learning specialists, data scientists, statisticians, information security analysts, and nurse practitioners. The action-oriented answer: invest in skills the data shows are growing. The site does not provide individual career advice; it points to the cited data at /whats-growing/.

Per LinkedIn data published April 2026, hiring is down approximately 20% since 2022. LinkedIn says AI is not yet the demonstrable cause. Per workforce trackers, several thousand 2025 job cuts in the US were directly attributed to AI by employers. The displacement is real but smaller than aggregate headline figures. Brookings 2025 (No AI Jobs Apocalypse, For Now) is the calibrated reference.

The OECD AI Working Group and the ILO have produced complementary work measuring per-occupation exposure to AI. The ILO 2025 refined Generative AI and Jobs index assesses ISCO-08 6-digit occupations across approximately 30,000 tasks and assigns each occupation a four-band exposure gradient. It is the most-defensible per-occupation generative-AI exposure measure currently in the public domain.

Per WEF Future of Jobs Report 2025, the projection is 170 million new roles created by 2030 against 92 million displaced, for a net 78 million job increase globally. Per Harvard Business School Working Knowledge, generative AI creates new demand in augmentation-prone roles. Net-positive at the aggregate level; uneven at the occupation level. The transition is not symmetrical, which is documented at /whats-growing/.

The calculator gives the ILO 2025 exposure gradient for the matched occupation, the Brookings 2024 task-level breakdown, and the BLS 2024-2034 projection. The honest answer: occupation-level exposure is a probabilistic distribution, not a personal prediction. The site does not predict your individual job; it reports what the cited primary sources publish. The full caveat is at /how-to-argue-with-this/.

Exposure is the share of tasks in an occupation that AI can technically perform. Displacement is the actual labour-market outcome (workers laid off, roles eliminated). They are not the same. Exposure is high in many occupations where displacement has been low to date. The distinction underpins this calculator's methodology and is defined in the glossary at /glossary/#ai-exposure and /glossary/#ai-displacement.

Per BLS 2024-2034 and WEF 2025, the projected-fastest-growing AI-adjacent roles include data scientists (BLS projects 34% growth), software developers (17%), information security analysts, AI and machine-learning specialists, and statisticians. Frontline jobs (healthcare, education, skilled trades) are also projected to grow on demographic grounds. Specific BLS figures are at /whats-growing/.

Per Brookings 2025 (New Data Show No AI Jobs Apocalypse, For Now), aggregate-labour-market data through mid-2025 does not show AI-driven mass unemployment. The disruption is real, task-level, and uneven. The apocalypse framing is journalistic; the supporting data does not require it. The calibrated framing is exposure-without-mass-displacement, which is documented at /how-to-argue-with-this/.