The future, with
people at its centre.

Large-scale displacement remains limited so far. The stronger near-term evidence is uneven time saving, changing work organisation and pressure on opportunities for younger workers. That is not reassurance to do nothing; it tells us where to watch and respond first.

Yes—but not necessarily the same tasks, entry routes or conditions. The ILO estimates one in four workers is in an occupation with some GenAI exposure and says transformation is more likely than full replacement because human input remains necessary.

The warning deserves attention, but the broad evidence is still mixed. OECD analysis did not find a sustained, disproportionate fall in junior postings after widespread LLM use began. Some occupations and younger workers may still face real pressure, so AIm watches hiring, training access and job quality instead of treating one headline as a universal trend.

The strongest direction is not simply 'learn to code'. AI adoption is increasing the value of higher-order thinking, communication, adaptability, human agency and practical digital confidence. Specialist AI roles are growing, but they remain a small part of the labour market. Most people need to combine useful AI literacy with knowledge of their own work and responsibility for the outcome.

AI enters a labour market that already faces a much larger jobs challenge: about 1.2 billion young people are expected to enter the workforce in emerging markets over the next decade, while roughly 400 million jobs are currently projected. AI could help people and firms create new value, or weaken starter pathways if it is used only to remove cost. The outcome depends on skills, entrepreneurship, infrastructure and whether employers keep ways for people to learn.

It could protect lower-income households, but it has funding and economic trade-offs and is not the only way to share AI gains. This is one model, not a forecast. Policy also has to consider services, wages, work opportunities, ownership and the cost of living.

AI infrastructure spending is supporting near-term growth, especially in economies supplying chips and hardware, but investment demand is not the same as realised productivity or shared benefit. The IMF warns that gains remain uneven and that debt-financed projects, circular financing and disappointing returns could cause sharp reversals. AIm watches wages, time, prices, services, local infrastructure and ownership as well as headline growth.

No. The World Bank estimates current jobs in high-income countries are more than three times as likely to face GenAI automation risk as jobs in lower-income countries. Developing economies may still miss the productivity benefit when electricity, connectivity, skills or institutions are weak. Country, occupation and access matter more than one global headline.

South Africa has high unemployment alongside unfilled digital roles. The immediate opportunity is not to train everyone for one fashionable AI title; it is to strengthen foundational skills, connect learning to real employer demand and create more work-integrated learning. Software development and cybersecurity remain useful pathways, while every field increasingly benefits from practical digital confidence and supervised experience.

Yes. Across 135 countries, the research finds that workers in exposed jobs may already be connected to the tools that can automate tasks, while weak electricity, connectivity, skills and institutions can delay the wider productivity benefit. For developing economies, the danger is not only job loss; it is disruption without enough new capability, investment or shared gain.