640,000 New Jobs Later, AI Still Gets Blamed for Everything

640,000 new roles created between 2023 and 2025. That’s what AI has actually done to the job market, according to recent data. Yet somehow, we’re still having the same tired conversation about AI stealing jobs.

I’ve spent the last year testing AI toolkits for this site, and I’ve noticed something interesting: the tools that actually work don’t replace people. They make existing jobs more tolerable. The ones that promise to “automate your entire workflow” usually end up creating more work than they save.

So when I see LinkedIn data showing that only 9% of companies have actually replaced jobs with AI, I’m not surprised. What does surprise me is that 60% of hiring managers are still blaming AI for their hiring freezes. That’s some impressive mental gymnastics.

The Real Story Behind the Numbers

Tech job openings have rebounded sharply in 2026. This directly contradicts the narrative that AI is wiping out engineering roles. But narratives are sticky, especially when they give us something convenient to blame.

Here’s what I think is actually happening: companies overhired during the pandemic boom. They’re now correcting that mistake. AI makes for a much better scapegoat than admitting “we got our growth projections wildly wrong and hired 3,000 people we didn’t need.”

The data backs this up. If AI were truly replacing human workers at scale, we wouldn’t be seeing job creation in the hundreds of thousands. We’d be seeing the opposite.

What AI Actually Does

From my testing experience, AI tools fall into three categories:

• Tools that save you 20 minutes a day on repetitive tasks
• Tools that create new work by requiring constant supervision and correction
• Tools that enable you to do things you literally couldn’t do before

That third category is where the job creation happens. When you can suddenly analyze datasets that were previously too large, or generate variations on designs that would have taken weeks, you’re not replacing work. You’re expanding what’s possible.

The LinkedIn data suggests AI is “amplifying, not replacing, human roles.” That matches what I see in the toolkit space. The best AI products don’t eliminate the need for human judgment. They give humans better tools to exercise that judgment.

Why the Fear Persists

So why does everyone keep predicting mass unemployment? Partly because doom sells. Articles claiming “80% of jobs will be gone in 3 years” get more clicks than “AI creates new job categories.”

But there’s also a kernel of truth in the fear. AI will change what jobs look like. Some roles will disappear. Others will emerge. That transition period is uncomfortable, and discomfort breeds anxiety.

What frustrates me as someone who tests these tools daily is how much of the conversation happens in the abstract. People debate whether AI will replace developers without ever actually using AI coding assistants. They predict the death of creative work without testing AI design tools.

The Hiring Freeze Reality

When 60% of hiring managers cite AI as a reason for hiring slowdowns, but only 9% have actually replaced positions with AI, there’s a disconnect. That gap tells you more about corporate messaging than about AI’s actual impact.

Companies are using AI as cover for normal business cycle corrections. It sounds better in a board meeting to say “we’re strategically repositioning for an AI-first future” than “we hired too many people and now we need to cut costs.”

The tech hiring rebound in 2026 suggests the market is correcting itself. Companies that paused hiring are starting again. The sky didn’t fall. Engineers still have jobs. In fact, there are more AI-related roles than ever before.

What This Means for You

If you’re worried about AI taking your job, the data suggests you should be more worried about normal economic cycles and corporate decision-making. AI is creating more roles than it’s eliminating, at least for now.

But that doesn’t mean you can ignore AI. The roles being created require AI literacy. You don’t need to become a machine learning engineer, but you do need to understand how to work alongside these tools.

The job market isn’t being destroyed by AI. It’s being reshaped by it. Those are two very different things, and the distinction matters.

🕒 Published: April 16, 2026