AI's clearest effect on the job market is the thinning of the bottom rung. Why entry-level roles are shrinking, which are most exposed, and what starters should do.
The most visible effect of AI on the job market is not mass unemployment. It is the quiet hollowing out of the bottom rung. Entry-level roles, the jobs that absorbed graduates and career starters and taught them the trade, are thinning, because the work those jobs were built on is the work AI does best. This is not a temporary dip in a hiring cycle. It is a structural shift in how the bottom of the ladder is built, and understanding why it is happening is the first step to navigating it instead of being caught by it.
Entry-level work was mostly the substitutable layer
Every role is a stack of two layers: routine production that follows known patterns, and accountable judgment that does not. Entry-level jobs were deliberately weighted toward the first. That was the point. You gave a junior the research, the first drafts, the data entry, the reconciliation, the routine processing, both because the organisation needed it done and because doing it was how the junior learned. The trouble is that this routine layer is precisely what AI now produces quickly and cheaply. The economic reason to hire someone to do it has weakened, and with it the traditional entry role.
The broken apprenticeship
The deeper problem is not just fewer openings. It is that the entry layer was the apprenticeship. Doing the routine work was how people built the judgment that the durable layer requires. If AI does the routine work, the rung that taught the trade is missing, and the path from beginner to expert loses its first few steps. Organisations have not solved this yet. Many are simply hiring fewer juniors and leaning on AI plus existing senior staff, which is efficient in the short term and quietly corrosive in the long term, because it stops producing the next generation of experts.
Which entry points are most exposed
The hollowing out is uneven. Entry roles that were almost entirely routine production are the most exposed: junior roles built around drafting, formatting, basic research, data entry, and routine processing. Entry roles that involve early accountability, real human contact, or hands-on physical judgment are far more resilient. The pattern across which jobs AI is replacing holds at the junior level too: it is the routine layer that goes, and entry roles that were mostly that layer feel it first and hardest.
- More exposed: junior analyst roles centred on data pulls and standard reports, entry content roles centred on first drafts, routine administrative processing.
- More resilient: roles with early client or patient contact, roles with hands-on judgment, roles where a junior owns a small but real decision from the start.
- The dividing line is whether the entry job gives you any accountable work, or only routine production.
What career starters should do differently
If the routine entry job is thinning, the move is to aim at entry points that put you near durable work sooner, even if they are harder to get. Prefer a role where you own a small decision over one where you only produce output. Bring proof of judgment, not just a credential, because polished output is now cheap and demonstrated ownership is not. And use AI tools to clear the routine work fast so you can spend your early career building the judgment the tools cannot supply. The goal is to compress the broken apprenticeship by deliberately seeking the learning the old routine rung used to provide. For more on which starting roles still make structural sense, see the guidance on entry-level jobs and AI risk.
What this means for employers
Organisations that respond to AI by simply cutting junior hiring are mortgaging their own pipeline. The routine work that juniors did can go to AI, but the development that came with it has to be rebuilt deliberately, by giving new hires accountable work earlier and using AI to handle the production that used to fill their days. The companies that figure out how to grow expertise without the old routine apprenticeship will have a durable advantage, because expertise is exactly the scarce input the AI economy rewards.
Read your starting position
If you are early in your career or about to enter the market, you do not have to guess how exposed your target entry role is. The AI Career Simulator lets you model a role you are considering and read its exposure and the share that is durable, so you can choose a starting point that still teaches you something the market will pay for. Comparing two entry options on structure, rather than on starting salary alone, is one of the highest-leverage decisions a career starter can make.
The bottom rung is being rebuilt, not removed. The opportunities that remain favour starters who aim at accountable work, prove judgment early, and use AI to skip the routine layer rather than compete with it. When you are ready to turn that into a concrete plan, the AI Career Index assessment reads your inputs and sequences the first moves, so starting out in a hollowed-out entry market becomes a navigable path rather than a closed door.
Run your structural assessment
The AI Career Index is free to take. Get your structural snapshot in 10 minutes, including your durability score, archetype, and projection.
Take the Free Assessment