Almost every AI jobs story in 2026 has been a story about loss, and there is no shortage of evidence for that frame. But the most cited forecast in the field, the World Economic Forum's Future of Jobs work, contains a second number that rarely makes the headline. By 2030, the WEF projects that AI and related technologies will create roughly 170 million new roles globally while displacing around 92 million, for a net gain of tens of millions of jobs. The disruption is real, but in aggregate the model points up, not down.
Churn, not collapse
The catch buried in that net figure is churn. A world that loses 92 million roles and gains 170 million is not a world where nothing changes. It is one where a vast number of people must move from shrinking occupations into growing ones, often requiring new skills, sometimes in new places. The optimism in the headline number depends entirely on whether that transition actually happens, and the early signs in graduate and entry level hiring suggest the on ramp is narrowing even as the destination roles multiply.
The premium the market is paying
Where demand is real, pay is following. Analysis of more than 10 million job postings in the United Kingdom found that candidates advertising AI related skills command, on average, a salary around 23 percent higher than otherwise comparable peers without them. Separate research from labour market analytics firm Lightcast has tracked a similar premium across markets, and the WEF has pointed to evidence that AI skills can offset disadvantages of age and conventional education in the hiring process. The signal to workers is blunt, capability now beats credential.
The supply problem
The constraint is not demand but supply. Training pipelines have not kept pace with the surge in employer appetite, leaving a genuine shortage of qualified candidates and a widening gap between what firms want and what the workforce can offer. That shortage is precisely what sustains the wage premium, and it is also what makes the WEF's net job creation scenario fragile. Roles that go unfilled because no one is trained for them do not show up as the new jobs the forecast promises.
Reading the two numbers together
The honest version of the AI jobs story holds both figures at once. The 92 million displaced explains the anxiety, the layoffs and the booed commencement speakers. The 170 million created explains why reskilling, not resignation, is the rational response. The risk in 2026 is that the loss arrives first and concentrated, on the young and the junior, while the gains arrive later and diffuse, leaving a real and painful gap in between even if the long run arithmetic is benign.
