The headline figures are real. The interpretation usually isn't.
Barely a week goes by without a new report on how many jobs AI will eliminate. The numbers are alarming by design. They're also frequently taken out of context.
Let's look at what the leading research actually says.
The World Economic Forum's Future of Jobs Report 2025 projects that by 2030, approximately 92 million roles will be displaced by AI and automation. In the same report: 170 million new roles will be created. Net result on paper: +78 million jobs globally.
But here's what most articles skip: the WEF itself notes a net displacement of approximately 14 million jobs (roughly 2% of global employment) when you account for transition friction, skills mismatches, and geographic concentration. The math doesn't add up cleanly because the losses and gains don't happen in the same place, at the same time, or for the same people.
Goldman Sachs puts it differently: generative AI could automate tasks equivalent to 300 million full-time jobs worldwide, with two thirds of current occupations exposed to some degree of automation. That's not the same as 300 million people losing their jobs. But it does mean that two thirds of the workforce will see their daily work change, substantially, within this decade.
McKinsey's research puts the automation potential of current work activities at 60 to 70% before 2030. Not jobs. Tasks. The distinction matters enormously, and most public discourse collapses it.
The Sparagus read: The question isn't "will AI take my job?" for most people. The more accurate question is "will AI change what my job requires, and am I positioned to evolve with it?" For organisations, the equivalent question is: "Are we building the capability to integrate these tools, or are we waiting for the dust to settle?" The dust isn't settling. It's accelerating.
What employers are actually planning (and it's more nuanced than the headlines)
41% of employers globally plan to use AI to reduce headcount, according to the same WEF report. That number gets quoted a lot. Here's the number that gets quoted less:
77% of those same employers plan to upskill staff to work alongside AI. And 47% say they intend to move affected employees into different roles internally rather than letting them go.
So the dominant employer strategy isn't replacement. It's redeployment, combined with selective reduction in roles that are genuinely becoming redundant.
The problem is execution. According to BCG's AI at Work 2025 report, around 2 in 3 executives say that generative AI adoption has led to tension and division within their organisation. 42% describe it as actively tearing their company apart.
That's not a technology problem. That's a change management problem.
Gallup's research identifies the single most important variable in whether AI adoption succeeds in an organisation: manager buy-in. When managers actively endorse and model AI use, adoption among their teams reaches 79%. Without that support, it drops to 34%. Nearly half the potential, gone, because of one layer of leadership.
The Sparagus read: Most organisations are not failing at AI because they chose the wrong tools. They're failing because they haven't made a genuine organisational commitment to integration. Adopting a licence is not a strategy. Without clear leadership from the top and genuine training investment throughout, AI becomes something people do quietly on the side, rather than something that transforms how the business operates.
Belgium specifically: third in Europe, but with a significant internal gap
The Belgian picture is one of genuine ambition combined with uneven execution.
According to PwC Belgium's Bridging the AI Gap report, a quarter of Belgian companies with more than 10 employees now use at least one AI application. That represents an 80% increase in a single year, and places Belgium third in Europe for AI adoption in business, behind only Denmark and Sweden.
That's impressive. Here's the gap:
- 40% of Belgian workers who use a computer regularly still use zero AI tools
- Regular use of generative AI has climbed from 13% to 34% in one year, which is fast by any standard
- 23% of Belgian companies expect to need fewer staff because of AI
- A third expect to need workers with different skills entirely
According to EY Belgium, 67% of the Belgian workforce has never heard of AI agents, which are rapidly becoming the primary deployment model for AI in enterprise environments. Meanwhile, 72% of employers in Benelux are already struggling to fill technical roles. The talent gap isn't coming. It's here.
The Sparagus read: Belgium is well-positioned relative to the rest of Europe, and that's genuinely good news. But position is not the same as readiness. The companies that will come out ahead aren't the ones with the most AI licences. They're the ones that have aligned their hiring strategy, their internal development programmes, and their leadership culture around this shift. Those that haven't started that process are not catching up in three months.
The employee side: anxious, underinformed, and more capable than employers think
According to ADP's global workforce survey, covering 39,000 workers across 36 countries, only 22% of employees worldwide strongly agree that their job is safe from AI. Among frontline workers, that number drops to 18%.
The anxiety is real. What's less well understood is that it coexists with genuine readiness. McKinsey's research found that C-suite leaders estimate only 4% of employees use AI for at least 30% of their daily work. The self-reported figure from employees is three times higher. Employees are ahead of where their managers think they are.
That gap is a symptom of something important: many organisations don't have visibility into how their people are actually using AI, because adoption is happening informally, unsanctioned, and in many cases, in silence.
The Sparagus read: The employee anxiety around AI is not primarily about the technology. It's about communication and trust. Workers who receive clear information about how AI will change their role, and what their organisation is doing to support the transition, are significantly more likely to engage constructively with it. The companies that treat this as a communication and culture challenge, not just a procurement decision, will retain better people through the transition.
In short
AI is not arriving in the future. It is restructuring work right now, at a pace that the standard corporate planning cycle was not built to handle. The headline numbers tell one story. The real story is about which organisations are moving with enough clarity and speed to capture the upside, and which are watching it happen from the outside.
For employers, the question is no longer whether to integrate AI. It's whether the integration is happening strategically or by accident. For talent functions specifically, the implications are significant: the roles being created in the next three years are not the same as the roles being filled today. Hiring strategies, skills frameworks, and development programmes built for 2022 will not deliver for 2026.
The speed of change isn't slowing down. The organisations that internalize that, and act accordingly, are the ones that will define what comes next.
