AI risks · 2030

Three scenarios for AI and work in 2030

Nobody knows how AI will affect work and society by 2030. The three scenarios below are illustrative, not predictions: they show a range of plausible paths, drawn from published scenario work and forecasts by the OECD, the International AI Safety Report, the IMF, the ILO and the World Economic Forum. Reality is likely to combine elements of all three, and to differ between countries, sectors and groups of workers.

These are illustrative scenarios, not predictions.

A

Gradual change: AI as a work tool

AI keeps improving, but more slowly than in recent years. By 2030 it works like a capable assistant that can handle many computer-based tasks but still needs people to direct and check it. Most jobs change rather than disappear, in line with the ILO's assessment. Employers' expectations in the WEF survey roughly hold: substantial churn, with clerical roles shrinking step by step, but a net increase in jobs overall as retraining keeps pace for most workers.

Signs this is happening

  • Employment in highly exposed occupations stays broadly stable, while their task content changes.
  • Most AI use is people working with AI rather than AI working alone.
  • Job ads increasingly ask for AI skills rather than disappearing.
  • Productivity rises moderately and steadily rather than in sudden jumps.

Based on: The OECD scenario in which progress slows, as presented in the International AI Safety Report 2026; the ILO's 2025 exposure index; and employer expectations in the WEF Future of Jobs Report 2025 (job churn of 22%, net growth of 7% by 2030).

B

Rapid automation of office and knowledge work

AI progress continues at today's pace or speeds up. By 2030 AI systems work like expert collaborators, or in the most extreme OECD scenario like remote workers at human level, and can carry out many digital tasks largely on their own. Clerical and entry-level knowledge work shrinks faster than people can retrain. In line with the IMF's warning, a large share of exposed jobs in advanced economies sees lower demand and wage pressure, even if productivity grows strongly. How the gains are shared becomes a central political question.

Signs this is happening

  • Hiring of young graduates and other early-career workers falls clearly in the most exposed occupations.
  • A growing share of AI use, especially by businesses, is full automation rather than assistance.
  • Companies increasingly cite AI as a reason for staff cuts or hiring freezes.
  • AI agents keep handling longer and more complex tasks, with the doubling time of roughly seven months holding or shortening.

Based on: The OECD scenarios in which progress continues or accelerates, as presented in the International AI Safety Report 2026, together with the report's account of economists who expect significant falls in wages and employment; the IMF's 2024 analysis of AI exposure in advanced economies.

C

Uneven change and plateau

AI capabilities stall, for example because of limits in data, energy or investment, and reliability problems mean systems still need close human support. Adoption stays patchy: high in some countries, sectors and large firms, low elsewhere. Some occupations, such as programming, change a lot while others barely notice. Productivity gains are modest and unevenly spread, which risks widening gaps between countries and between groups of workers rather than transforming the labour market as a whole.

Signs this is happening

  • Investment in AI and data centres slows or is scaled back.
  • AI use remains concentrated in a few tasks, such as writing code.
  • Productivity statistics show little overall change.
  • Large differences in adoption persist between big and small firms and between rich and poorer countries.

Based on: The OECD scenario in which progress stalls, as presented in the International AI Safety Report 2026, and the report's data on uneven adoption (over 50% of the population in some countries, likely under 10% in much of Africa, Asia and Latin America); the concentration of use in Anthropic's Economic Index; and the IMF's analysis of AI widening gaps between countries.

Sources

  1. International AI Safety Report 2026 (including OECD scenarios for AI progress to 2030) – International AI Safety Report (chair: Yoshua Bengio), 2026
  2. The Future of Jobs Report 2025 – World Economic Forum, 2025
  3. AI Will Transform the Global Economy. Let's Make Sure It Benefits Humanity – International Monetary Fund (IMF), 2024
  4. Generative AI and Jobs: A Refined Global Index of Occupational Exposure (ILO Working Paper 140) – International Labour Organization (ILO), 2025
  5. Anthropic Economic Index report: Economic primitives – Anthropic, 2026

The risks of AI – what is real and what is hype?

A sober guide to the risks of AI: jobs, deepfakes, fraud, cybersecurity, bias, privacy, energy and loss of control – with sources and what you can do.

AI risks →