Every career used to start the same way: doing the boring stuff.
You summarized the meeting notes. You built the first draft of the deck. You cleaned the spreadsheet, checked the citations, answered the routine customer emails and sat in on calls you barely understood.
It wasn't glamorous. But it was how you learned. The grunt work was the apprenticeship.
Now AI does the grunt work. And nobody has figured out what replaces the apprenticeship.
The Bottom Rung Is Disappearing First
When companies talk about AI "efficiency," they rarely mean replacing senior experts. Senior people carry relationships, judgment and institutional knowledge that are hard to replicate.
They mean the tasks that are repetitive, screen-based and easy to check. In other words, the exact tasks we used to hand to new hires.
Junior analysts, research assistants, entry-level copywriters, first-line support, paralegals, junior developers. These roles share a profile: lots of routine work, supervised by someone more experienced. That makes them prime candidates for automation. I mapped the broader pattern in Which Jobs Will AI Replace First?
It doesn't usually look like mass layoffs. It looks like smaller graduate classes, fewer internships and job postings that ask for "two to three years of experience" for roles that used to be where you got that experience.
The Catch-22 Gets Worse
New grads have always faced the classic trap: you need experience to get a job, and a job to get experience.
AI tightens it. If a company can have a senior employee plus AI do the work that used to require a senior employee plus two juniors, the juniors never get hired. And if the juniors never get hired, the experience never gets built.
Add AI resume screening to the mix, and many applicants never even reach a human. (Here's how AI resume screening actually decides.)
Companies Are Eating Their Own Seed Corn
Here's the part executives should be losing sleep over.
Every senior expert was once a junior employee who made mistakes, got corrected and slowly developed judgment. Cut the junior roles, and in five or ten years you have a shortage of mid-level and senior talent, because nobody was trained.
Translation: automating the bottom rung saves money this quarter and creates a talent crisis next decade.
It also concentrates knowledge dangerously. When only a handful of experienced people understand how things actually work, and AI does the rest, organizations become brittle. Who catches the AI's mistakes once those experts retire? (This is the same problem behind companies losing money on unvetted AI.)
The Irony for Tech Grads
For years, the safe advice was "major in computer science." Now coding assistants handle a growing share of routine programming, and entry-level developer roles are among the most exposed. The builders are automating their own on-ramp, a pattern I wrote about in The People Building AI Are Automating You Out of a Job.
So Where Do New Grads Start?
The ladder isn't gone. It's been rearranged. Here's how to climb the new one.
Become fluent in AI, then go one level deeper. Employers expect you to use the tools. What sets you apart is understanding the work well enough to judge the output. Be the junior who catches the mistake, not the one who forwards it.
Build proof, not just credentials. A portfolio of real projects, case studies or published work shows ability in a way a degree alone no longer does. Show what you made and how you thought about it.
Pick roles with human contact. Client-facing, operations, sales, field work and healthcare roles keep a stronger human core and still offer real learning.
Target smaller companies. Startups and small businesses often can't afford to be picky, and you'll touch more of the business faster.
Network like it matters, because it does. Referrals bypass algorithmic filters. A conversation with a real person is worth a hundred online applications.
Protect your thinking. Use AI to move faster, but do the reasoning yourself, especially early in your career, when you're building judgment. Outsource it now and you may never develop it. (See Is AI Making Us Dumber?)
For a longer playbook, read How to Future-Proof Your Career Against AI.
What Companies Should Do Instead
Let's be clear: this isn't just a problem for graduates to solve on their own.
Companies that keep hiring juniors, and redesign those roles around learning instead of grunt work, will have the experienced talent everyone else is scrambling for in a few years. Pair new hires with AI and with mentors. Give them real problems. Treat entry-level hiring as an investment, not a line item to automate away.
The ones that don't will discover an uncomfortable truth: you can automate tasks, but you can't automate experience.