The debate over whether AI destroys more jobs than it creates usually plays out in abstractions. LinkedIn's Economic Graph data puts a concrete number on one side of that ledger: AI has added more than 600,000 new AI-enabled data-centre jobs and 1.3 million new roles — including AI engineers, forward-deployed engineers, and data annotators — over the past three years, titles that largely didn't exist in the job market before.

That growth is happening alongside a much larger, more disruptive shift. The World Economic Forum projects 170 million new jobs and 92 million displaced globally by 2030 — a net gain of roughly 78 million, but one that comes with significant churn underneath it. The WEF also estimates that 22% of all jobs covered in its research will change structurally, that 39% of core job skills will shift, and that 41% of employers plan to cut headcount specifically in roles where AI can take over tasks.

Why both numbers can be true at once

Net job growth and widespread disruption aren't contradictory — they describe different things. The WEF's net-positive figure is an aggregate across the entire global economy; it says nothing about whether the people losing roles are the same people gaining the new ones. LinkedIn's own research found that AI-related conversation on its platform has surged 70%, and that more than half of its members are now in roles directly affected by AI in some way, with job skills for those roles expected to shift by 65% by 2030.

What the new roles actually look like

The job titles behind LinkedIn's 1.3 million figure cluster around a few patterns worth noting if you're planning a pivot:

• Infrastructure and operations — data-centre and AI-enabled infrastructure roles, which scale with compute demand rather than with any single company's AI strategy.
• Deployment and integration — “forward-deployed engineer” roles that sit between an AI vendor and a customer, translating a general-purpose model into something that works for one business's specific workflow.
• Data quality — annotation and evaluation roles that remain stubbornly human-dependent even as models improve, because someone still has to judge whether an output is actually good.

The practical takeaway

If your current role involves tasks that are easily specified and repetitive, the WEF's 41% figure is the one to take seriously — that's where employers say they expect to reduce headcount. If you're deciding where to invest your next round of upskilling, the roles behind LinkedIn's growth number point toward AI deployment, data infrastructure, and the evaluation work that sits around AI systems rather than inside them. Neither statistic on its own tells you what will happen to your specific job; together, they're a reasonable map of where the growth and the risk are each concentrated.