All

No Country for Young Grads? What the data actually says and what to do about it

All

No Country for Young Grads? What the data actually says and what to do about it

What the new Burning Glass reports show (July 2025)

“For the first time in modern history, a bachelor’s degree is failing to deliver on its fundamental promise: access to professional employment.” (Reference: The Burning Glass Institute)

1) Entry-level demand is falling where AI overlaps with junior work.
Job postings that require ≤3 years’ experience have dropped primarily in occupations with medium/high LLM exposure (e.g., marketing, project coordination, finance). In the same fields, demand for senior talent holds steady or rises.

2) The “Expertise Upheaval”: two different futures.

Growth Roles (e.g., legal associates, marketing specialists, project managers): AI automates foundational tasks; entry points narrow. ~18M workers (~12%). Pays more and more likely to require degrees.
Mastery Roles (e.g., network admins, data warehousing, loan interviewers): AI compresses time-to-proficiency; access can expand. ~29M workers (~19%).
Growth Roles pay ~28% more on average (≈ $20k), and are 51% more likely to require a degree—so the squeeze hits lucrative career ladders first.

3) It’s not only AI.
Lean post-pandemic staffing and manager risk-aversion are amplifying the effect (more preference for “ready now” hires).

4) A growing graduate surplus.
Class of 2023: 52% were underemployed one year out. And the number of college-educated workers is set to grow by 7–11M by 2034—intensifying competition unless entry routes change.

Important nuance: Entry-level opportunities have grown in low-exposure roles since 2022. The squeeze is targeted, not universal.

Methods note & uncertainty: Burning Glass maps learning curves using postings, wages, and AI task exposure. Timing and magnitude will vary by sector and adoption speed.

What the new Burning Glass reports show (July 2025)

“For the first time in modern history, a bachelor’s degree is failing to deliver on its fundamental promise: access to professional employment.” (Reference: The Burning Glass Institute)

2) The “Expertise Upheaval”: two different futures.

Growth Roles (e.g., legal associates, marketing specialists, project managers): AI automates foundational tasks; entry points narrow. ~18M workers (~12%). Pays more and more likely to require degrees.
Mastery Roles (e.g., network admins, data warehousing, loan interviewers): AI compresses time-to-proficiency; access can expand. ~29M workers (~19%).
Growth Roles pay ~28% more on average (≈ $20k), and are 51% more likely to require a degree—so the squeeze hits lucrative career ladders first.

3) It’s not only AI.
Lean post-pandemic staffing and manager risk-aversion are amplifying the effect (more preference for “ready now” hires).

Important nuance: Entry-level opportunities have grown in low-exposure roles since 2022. The squeeze is targeted, not universal.

Methods note & uncertainty: Burning Glass maps learning curves using postings, wages, and AI task exposure. Timing and magnitude will vary by sector and adoption speed.

So what? (tech × talent × futures)

For students & early-career pros

Aim for AI-complementary entry points: data-aware ops, client-facing coordination, QA/controls, analytics-lite roles. These build judgment + context (the scarce bits).
Build a dual stack: (1) tool fluency (prompting, retrieval, light scripting); (2) decision skills (scoping, verification, risk triage).
Treat “first drafts” as AI’s job; your job is specification, supervision, and integration.

For educators & training providers

Replace “capstone only” with progressive simulations that mimic missing junior tasks (client briefs, noisy data, red-team reviews).
Assess for directing AI (can learners decompose a task, set constraints, audit outputs?)—not just content recall.
Create bridge credentials tied to Growth-Role skills beyond first drafts: stakeholder comms, error-budgeting, and handoff quality.

For employers

Stop measuring pilots by demo wins; measure cycle time, error rates, rework, and customer outcomes.
Where entry-level rungs are eroding, build synthetic apprenticeships: structured rotations, sandboxed client sims, mentored reviews.
Widen aperture in Mastery-Role pipelines (adjacent-skill hiring; less degree filtering) while you protect and cultivate scarce Growth-Role expertise.

For policymakers & funders

Incentivize evidence-building: publish sector playbooks that specify tasks safe for AI vs. human oversight (with liability norms).
Co-fund on-ramp programmes that replicate practice (state–employer partnerships, outcomes-tied).
Track underemployment as a core KPI, not just unemployment, to see if pathways are working.

What to watch next

Posting mix: ratio of junior (≤3 yrs) to senior (6+ yrs) postings in Growth-exposed occupations. If the gap widens, rungs are still eroding.
Low-exposure roles: are they still absorbing entry-level demand—or flattening out?
Training substitutions: adoption of simulations/mentored reviews as replacements for missing junior tasks.
Graduate surplus trend: net growth of degree-holders vs. professional openings through 2030–34.

My take

Reframing is the unlock. Seeing AI through learning curves (not headlines) moves the debate from utopia/dystopia to design: where to create entry points, which skills to teach, how to share gains.
If you’re hiring, teaching, or job-seeking: which entry point would you build—or rebuild—first?

Sources (July 2025):
Burning Glass Institute, No Country for Young Grads (report + overview). The Burning Glass Institute
Burning Glass Institute, The Expertise Upheaval (report + overview). The Burning Glass Institute