Every other week there’s a new headline. AI is Going to Take Over Millions of Jobs. AI is Making More Jobs Than It Steals. Half of Office Workers Will Be Replaced by AI by 2027. It’s tiring, and let’s face it, it’s difficult to believe anything these days.
So what then, does AI take jobs away or does it make more jobs? The true answer, knowing the real statistics that will emerge in 2026, is — both. At the same time. But the whole story is quite more complicated than either of these headlines suggest.
The number Nobody Argues About
First, let’s begin with the source that most economists really believe: Goldman Sachs and the World Economic Forum’s Future of Jobs Report.
Currently, the most repeated word is AI: it may destroy 92 million jobs but at the same time create 170 million! Just do the maths, and that’s an increase of about 78 million jobs worldwide. Not a net loss. A gain.
The first time I saw that, I was surprised too, as it seems extremely at odds with the doom headlines. But it’s repeatedly been reported in several independent sources so it’s not an optimistic statistic cooked up for the ear.
Goldman Sachs also has its own estimate for the displacement of jobs by generative AI alone, estimating that it could eliminate 25 million full-time jobs by 2026 and up to 270 million by 2030. Until you realize the other side of that same report: Two-thirds of the current jobs are open to AI automation, but “exposed” does not equal “eliminated. Generally it is the word “changed”.
So Which Jobs Are Actually Disappearing?
This is where it becomes more specific — and where the real worry lies — as it’s not distributed evenly.
The most affected industries are administrative (26% of jobs affected), customer service (20%) and production (13%), according to McKinsey’s analysis. The legal sector and the education sector are only slightly affected, with a 6% effect in both professions. The fewest jobs are impacted by management roles (3%).
The top countries in terms of exposure are, by country, actually the richest: Switzerland (71%), South Korea (70%), Japan (68%) and the UK (67%). Approximately 59% of the jobs in the United States are exposed to some degree of AI automation. This is the very opposite to what many people would think. AI is not threatening low-income countries primarily, it’s taking a shot at advanced economies, which have more to play to be automated by AI — more white collar, computer-based work.
As of 2026, 20% of companies expect to have an organization with a single-level hierarchy by the end of 2026, which may include a significant number of job cuts in middle management roles among these companies. That’s a number that middle managers should take note of.
And What Jobs Are Actually Being Created?
But, the one thing that doesn’t get nearly enough attention: “New jobs created” is not as good of a clickable headline as “jobs destroyed.”
AI Engineers have seen more than a 140% increase in growth, ranking among the fastest-growing job categories available today. AI Content Creator jobs have increased by more than 130%. The demand for roles such as Prompt Engineer, AI Solutions Architect, and AI Product Manager is expanding at various rates, ranging from 35% to 110% growth per job title.
Niche jobs are no more. Just as “Social Media Manager” became a standard job title 10 years ago, they’re becoming the norm in hiring.
There is also a wage effect, I believe which is the most salient number in this entire topic: Workers that can prove their AI skills now make up to 56% more than workers with the same skill set who can’t. In fact, that gap was 25% only two years ago. It grew by 100% over 2 years.
Why The Fear Bigger Than the Data Suggests
So why is all the world so anxious if the net numbers are positive?
Some of it has to do with timing. The jobs that are being lost are being lost right now, and that’s the ones in real companies, to real people. The jobs that are being created take time to materialize, and fall into new job categories that haven’t yet been cataloged in people’s mental lexicon of “real jobs.”
Partly it is a legitimate worry about certain groups. Multiple surveys in 2026 show that one in four workers globally is concerned that AI will replace their jobs — and it’s significantly more common among workers of color and among younger workers. Specifically for creative work, 81.6% of digital marketers think that AI will end their roles as content writers. That’s not paranoia. That is a fair interpretation of what’s going on in that niche at this time.
There’s also the generational change that’s on the verge of being overlooked. Approximately 40% of young university graduates selected a career path in a trade such as plumbing, construction and electrical work because it’s harder to automate those occupations.
What This Actually Means for You
If you are reading this because you are concerned about your own job, here is the actionable lesson we’ve learned from all the above:
The jobs in danger are process-oriented, repetitive and don’t involve much judgment or the need to interact with humans, like data entry, simple customer service scripts, and routine administrative tasks. The most rapidly expanding jobs are not those that are going to replace AI, but rather those that are going to work with AI and under its direction — and the wage premium for these skills is every bit as real and as fast as it gets bigger.
The truthful, yet no less unsexy guidance is to give this: The ability to effectively engage with AI in your own industry is now valued more than ever before, even just a year ago. Not due to safety, but because the data is quite clear that those working with AI are getting the financial edge.
The Real Answer
So, does AI take jobs or replace them with new opportunities? It will do both, and at the moment, it lets the creation side win by a considerable margin with each of the major reports from 2026. That is no small number of new jobs versus lost jobs, 170 million to 92 million.
But, if the average is positive, it does not imply that each one is okay. If you’re in a job that is changing to incorporate AI, but not being replaced by AI itself, you are probably on the winning side of this shift in work, and the wage data reflects that.
The news headlines will continue to report this in a very abbreviated manner, both sides of the story. The true history is more complicated and more uneven, and — if the majority who are willing to adapt are willing to admit it — more hopeful.