There's a version of the AI jobs story that gets all the airtime: mass layoffs, robots replacing humans, apocalyptic headlines. But the more interesting, more honest version is quieter. Companies adopting AI aren't firing people in droves. They're just not hiring the next person. No dramatic layoff announcements. No restructuring memos. Just a role that used to get backfilled, quietly left open. A team of five that used to become a team of seven, staying at five. A junior position that never makes it off the hiring plan. This is the hiring slowdown nobody wants to talk about, and the data is starting to show it clearly.

What the research actually says

A landmark study from Stanford's Digital Economy Lab, using ADP payroll data covering millions of workers, found that employment for 22 to 25 year olds in AI-exposed occupations fell 6 percent between late 2022 and July 2025. In contrast, employment among workers aged 30 and older in those same roles grew between 6 and 13 percent. The pattern is sharper when you zoom into specific roles. Employment for the youngest software developers was 20 percent below its late 2022 peak by mid-2025. For early-career customer service workers, the drop was nearly 11 percent. Meanwhile, employment for health aides, whose jobs have little AI exposure, was up across all age groups. Anthropic's own labor market research echoes this. Using a new measure combining theoretical AI capability with real-world usage data, they found "suggestive evidence that hiring of younger workers has slowed" in exposed occupations, with a 14 percent drop in the job-finding rate for workers aged 22 to 25 in AI-exposed roles compared to 2022 levels. Critically, they found no such decrease for workers older than 25. A Morgan Stanley survey across five sectors found a 4 percent net reduction in jobs among companies most exposed to AI adoption, with cuts "more pronounced in larger corporations" and affecting "mostly entry-level employees." At the same time, productivity increased 11.5 percent on average. And this isn't just about companies that have already deployed AI. According to a Harvard Business Review survey, 39 percent of organizations have already made low-to-moderate headcount reductions in anticipation of AI, with another 21 percent making large reductions. A further 29 percent are simply hiring fewer people than normal, waiting to see what AI can handle first. Only 2 percent of organizations reported large headcount reductions tied to actual AI implementation. That last number is the one worth sitting with. Most of the hiring slowdown isn't happening because AI proved it could do the job. It's happening because leaders believe it will.

This is not the same as AI layoffs

It's important to distinguish this from the wave of "AI layoffs" that made headlines over the past two years. Many of those were traditional cost-cutting exercises dressed up in AI language, companies using the AI narrative to justify decisions that were really about margins, restructuring, or investor signalling. The hiring slowdown is different. It's not about replacing people who are already there. It's about never bringing in the people who would have been next. The framing matters: AI isn't replacing your job. It's making your team productive enough that the next hire keeps getting deferred.

Who gets hit hardest

The pattern is consistent across every dataset: junior roles, support roles, and roles heavy on repeatable tasks are bearing the brunt. An IDC survey found that 66 percent of enterprises are reducing entry-level hiring due to AI. A Korn Ferry report found that 37 percent of organizations plan to replace early-career roles with AI entirely. The World Economic Forum's Future of Jobs Report 2025 found that 40 percent of employers expect to reduce their workforce where AI can automate tasks. The Indeed Hiring Lab paints the divergence vividly. At the end of 2025, job postings mentioning AI were 134 percent above their February 2020 baseline. Total job postings? Just 6 percent above. AI is creating demand for a specific kind of worker while the broader hiring market stagnates. Stack Overflow reported that the unemployment rate for workers aged 22 to 27 sits at 7.4 percent, nearly double the national average. Entry-level tech hiring decreased 25 percent year-over-year in 2024. This isn't just about coding. Customer service, data entry, financial analysis, content production, and administrative roles are all seeing the same pattern. The common thread is that these roles involve tasks that current AI tools can handle reasonably well, especially when supervised by a more experienced person.

The amplifier effect and the steeper ramp

Here's the dynamic that makes this genuinely tricky: AI functions as an amplifier. Experienced professionals with deep domain knowledge are becoming significantly more productive. A senior developer who knows the codebase, understands the architecture, and has years of context can use AI tools to move faster than ever. A senior analyst who knows which questions to ask can use AI to process data at a pace that would have required a team. But that amplification works precisely because these people already have the judgment, context, and expertise to direct the tools effectively. For someone just starting out, AI tools don't amplify much because there isn't much to amplify yet. This creates a paradox. The same technology that makes experienced workers more productive makes the entry ramp for new workers steeper. Companies need fewer juniors because their seniors are more capable, and the juniors who do get hired need to operate at a higher baseline from day one. IBM noticed this trap and reversed course. After initially leaning into AI-driven efficiency, the company announced it's tripling its Gen Z entry-level hiring. IBM's HR chief argued that cutting junior headcount risks creating an eventual shortage of mid-level managers, and that attempting to poach talent later will be costlier than building it now.

New roles are emerging, but the skills gap is real

It's not all contraction. New roles are appearing: AI orchestration, prompt engineering, agent operations, AI safety, model evaluation. The Indeed data shows that AI-specific job postings are at record highs. But these roles require fundamentally different skills from the entry-level positions they're displacing. A graduating computer science student trained in traditional software development doesn't automatically qualify for an AI operations role. The transition isn't seamless, and the retraining pipeline barely exists. The World Economic Forum estimates that 39 percent of current workers' skills will become outdated by 2030. PwC's Global AI Jobs Barometer, analysing close to a billion job ads across six continents, found that AI can make people more valuable, but only if they have the right skills to work alongside it. The gap between "AI-exposed" and "AI-ready" is where a lot of people are getting stuck.

A note on alternative explanations

Intellectual honesty requires acknowledging that AI isn't the only factor. Axios pointed out that the deterioration in the job market for young workers began before ChatGPT was widely available, coinciding with the Federal Reserve's aggressive monetary tightening in 2022 and 2023. J.P. Morgan Asset Management argues that less than 10 percent of firms actually use AI to produce goods and services today, and that today's cautious hiring environment reflects "more fundamental macro drivers: policy uncertainty, constrained labor supply and slowing economic growth." Schroders' analysis noted that outside the US, there has been no discernible acceleration in labour productivity, suggesting AI's impact on job markets beyond America has been limited so far. These are fair points. The hiring slowdown isn't purely an AI story. But the research increasingly shows that AI is a meaningful contributing factor, especially for specific demographics and role types, and the anticipatory effect alone (companies cutting before AI even delivers) is significant enough to warrant attention.

What this means if you're starting out

If you're a student or early-career professional, especially in Singapore or other knowledge-economy hubs, this is worth paying attention to without panicking over. The structural reality is that the traditional career ladder, where you start as a junior doing well-defined tasks and gradually take on more complexity, is getting compressed. Companies expect new hires to contribute at a higher level sooner, often using AI tools as a baseline rather than a bonus. This doesn't mean "just learn to code with AI" and everything will be fine. That framing dismisses the genuine structural challenge. The entry-level roles that used to teach you how an organisation works, how decisions get made, and how to collaborate, those roles are thinning out. And no AI tutorial replaces that kind of learning. What it does mean is that building judgment, context, and domain expertise early matters more than ever. Side projects where you make real decisions, not just follow instructions. Internships where you're embedded in actual workflows. An understanding of why things are done, not just how. The people who will thrive are the ones who treat AI as a tool that requires human judgment to direct, and who build that judgment as quickly as possible.

The honest take

This isn't doom. AI is genuinely making people more productive, and in the long run, productivity growth is how economies create prosperity. The World Economic Forum projects AI will create 11 million jobs while displacing 9 million others. The net is positive, at least in aggregate. But pretending the transition isn't happening is irresponsible. The costs are real and they're concentrated on the people least equipped to bear them: young workers, career changers, and anyone whose skills sit squarely in the automation sweet spot. Employee anxiety about AI-related job loss has jumped from 28 percent in 2024 to 40 percent in 2026, according to Mercer's Global Talent Trends report. That anxiety isn't irrational. It's a reasonable response to a labour market that is quietly, measurably shifting. The companies that get this right will be the ones that invest in building their talent pipeline even as AI makes their current teams more efficient. The ones that get it wrong will find themselves, a few years from now, with a hollowed-out middle layer and no bench strength. And the rest of us need to stop pretending this is either the apocalypse or a non-event. It's neither. It's a slow, real, structural shift, and it deserves a clear-eyed conversation.

References