73% of under-30s now expect AI to cut jobs as entry-level hiring collapses

Seventy-three percent of Americans under 30 believe artificial intelligence will eliminate jobs, up from 61% two years earlier, according to Pew Research Center data published August 18, 2026. That anxiety now mirrors a measurable shift in the labor market: entry-level vacancies are shrinking, and the roles that remain look nothing like the ones a decade ago.

The Pew figure, drawn from a short-read survey of young adults, is the clearest signal yet that graduates are entering a hiring environment reshaped by generative AI. A December 8, 2025 CNBC analysis documented the contraction in entry-level postings. PwC's AI Jobs Barometer added a sharper finding: in occupations most exposed to AI, junior hires are seven times as likely to require midcareer capabilities as those placed in less AI-connected roles.

What is "seniorization" of junior roles?

A June 18, 2026 Fortune report labeled the trend "seniorization" — the displacement of tasks that once defined a junior's first years. Drafting reports, analyzing data sets and taking meeting notes are now automated. New hires spend their days reviewing AI-generated output rather than producing work from scratch, a shift the analysis describes as "learning by reviewing" instead of "learning by doing."

The result is a paradox at the heart of the early-career pipeline. Evaluating AI output requires judgment, pattern recognition and critical thinking — skills that take years to build. Those skills are now exactly what employers want from day-one hires, even as the entry-level rung of the ladder disappears.

Junior employees are the most exposed to AI-generated errors and "authoritative slop," the analysis argues, because they have the least experience distinguishing useful model output from confident nonsense. Senior staff can also over-rely on AI, but the exposure is highest where the experience base is thinnest.

What should leaders do about the soft-skills gap?

A new analysis lays out three prescriptions for closing what it calls a growing early-career soft-skills gap. The recommendations are aimed at managers who want graduates to develop judgment while still working alongside AI.

• Run live, judgment-driven projects. Assign new hires projects they own end to end, then have them present to colleagues who can probe their reasoning in real time. The exercise is not a test. It is a structured opportunity to defend a recommendation in front of people who will challenge it.
• Make AI use visible and low-risk. Restrict early-career automation to a predefined set of simple tasks, such as summarizing an internal call or condensing a data set. Log every step in a shared document and require colleagues to flag AI-assisted content as it moves through review. The log becomes an accountability trail that lets managers catch misses early.
• Start AI-fluency training on day one. Treat fluency as more than prompt craft. Juniors need to know when AI should and should not be used, which tool fits which task, and how to critique model output. Pair the training with real organizational case studies that show where AI workflows worked and where they failed.

Why does this matter beyond the class of 2026?

If the pipeline keeps narrowing, the consequences compound. Employers get a thinner pool of mid-level candidates five years from now, because fewer graduates get the reps today. Graduates, in turn, face a Catch-22: jobs that once taught the craft now demand the craft as a prerequisite.

The analysis is blunt about the line companies should not cross. "AI isn't a shortcut for thinking," it warns, "nor is it a substitute for effort." Firms that bake that message into onboarding — and pair it with the kind of structured, judgment-heavy work described above — are most likely to keep developing the talent they will need in 2030 and beyond. Graduates who balance AI fluency with independent problem-solving, the argument goes, will be the ones leading teams by the end of the decade.