You Will Graduate Into a Different Job Market Than the One You Enrolled In
Planning in an uncertain economy
Picture a student arriving on campus in 2025. They pick a major, imagine a path, and trust that the world waiting on the other side of their degree will still resemble the one they know.
Two or three years later, the ground under that assumption has shifted. Sometimes I hear this directly from students when they drop by office hours or catch me after a talk. They feel the pull of change before they have the language for it. Artificial intelligence is already threaded into the routines that shape work and learning. It sits inside scheduling systems, applicant screeners, customer service platforms, and the software faculty rely on to run their courses. The tools show up quietly, which makes the transition easy to underestimate. A nursing student checks an AI tutor before lab. A small business owner adjusts a loan application with help from a chatbot. A department assistant watches the scheduling system complete work that used to take an afternoon.
These moments feel ordinary already, yet they signal something heavier. The job market students stepped into when they enrolled is not the one they will walk into after graduation. Programs, parents, and advisors will certainly still be working from older mental maps, and the mismatch will continue to grow a little each year.
The hardest part is how subtle the early changes look. Many employers do not announce major redesigns or layoffs. They hire fewer people, pause backfilling roles when staff leave (IBM, cited below), merge responsibilities across teams, or automate routine tasks that once trained new professionals (McKinsey, cited below). Junior work disappears first because AI systems absorb the predictable, repeatable tasks that entry-level workers used to learn on (OECD, cited below). When those tasks move upstream into automated workflows, the volume of true early-career work shrinks even if companies never run a single headline-grabbing layoff.
These quiet adjustments are already visible in job postings - in the case of software developers it’s been going on since fall of 2024 - that ask candidates to supervise AI-generated drafts rather than create the initial version themselves (NBER, cited below). They reflect a structural shift that students feel long before the data catches up.
We can all (parents, students, faculty) sense this tension long before we can name it. We worry about whether they picked the right major, whether the work they are preparing for will still exist, and whether they should learn every new tool they see online. Faculty and advisors feel their own version of this pressure. Course content ages even more quickly than it was. Program changes require governance, accreditation timelines, state approvals, and coordination across departments. Those processes exist to protect quality, yet they move slowly compared to how fast employers update workflows (McKinsey, cited below). Everyone wants to give students honest guidance, but the ground sometimes moves faster than the systems built to respond.
There is no single fix. Institutions face real limits: budgets that cannot stretch to rebuild every program at once, political constraints around curriculum changes, and the simple fact that faculty already work at capacity. Already on college campuses, we have had to sequence changes in a way that feels slower than the headlines, even when the urgency is obvious.
Still, there are responsibilities we cannot ignore. We owe students clarity. They deserve to know which roles are expanding, which are narrowing, and where new opportunities are forming. That kind of honesty does not discourage them. It steadies them.
We owe programs a chance to evolve. Some will need new layers of AI literacy, new emphasis on judgment and oversight, or cross-training that prepares students for work that still depends on human presence. Others will need a longer conversation because shifting them requires time, faculty agreement, and resources that colleges rarely have in surplus.
We owe adult learners realistic paths back into the classroom. Many displaced workers cannot simply step into AI-adjacent growth roles. Skills do not always transfer cleanly, interests do not always align with technical work, and life constraints such as caregiving or transportation limit what is possible (Aspen Institute, cited below). Geography matters because new jobs do not appear evenly across regions. Reskilling works, but it is uneven, slow, and resource-dependent (Brookings and WEF, both cited below).
We owe employers a partnership shaped by candor rather than vague optimism (default response from those most affected). Many are navigating their own confusion about tools, workflows, and hiring. When they share what they are seeing, it helps programs stay aligned with the work students will encounter.
And we owe ourselves the space to acknowledge that AI operates at a system level most people never see. Research shows that many workers cannot tell whether their employer already uses AI behind the scenes (MIT Sloan, cited below). That invisibility contributes to a sense of instability long before workers experience direct change.
And we owe ourselves the space to admit that some questions remain unsettled. There are parts of this transition that move faster than our planning cycles, and occasionally faster than our imagination. Naming that uncertainty is better than promising stability we cannot guarantee.
Students entering college today will graduate into a labor market with unfamiliar contours, and the early signals are visible right now. Job postings are shifting (NBER, cited below), entry-level opportunities are stretching thinner (McKinsey and OECD, cited below), and AI-enabled screening systems are filtering candidates in ways most applicants never see. Workers often feel these changes emotionally before the data reflects them, because freezes, consolidations, and rising expectations change the texture of opportunity long before unemployment numbers move (Aspen Institute, cited below).
This is a call to pay attention, not panic - although a little panic might not hurt right now. It asks institutions to be more flexible, more transparent, and more willing to update old assumptions. It asks students to develop a kind of alertness that earlier generations did not need, because the path ahead of them is being redrawn as they walk it.
If we meet them with honesty and a plan that reflects the world as it is, their chances of building a sturdy future grow stronger. They do not need perfection from us, and we will not solve every part of this shift at once. What they need is a community willing to face the uncertainty with them and to keep moving, even when the way forward feels partly unfinished.
Citations
IBM hiring and backfilling pause
IBM Considers Replacing 7,800 Jobs With AI
https://www.bloomberg.com/news/articles/2023-05-01/ibm-ceo-says-thousands-of-back-office-jobs-could-be-replaced-by-ai
Entry-level role contraction and task automation
McKinsey: Generative AI and the Future of Work in America
https://www.mckinsey.com/featured-insights/future-of-work/generative-ai-and-the-future-of-work-in-america
OECD Employment Outlook
https://www.oecd.org/employment-outlook/
NBER: AI and Hiring Effects
https://www.nber.org/papers/w31874
Reskilling constraints
Aspen Institute Future of Work Initiative
https://www.aspeninstitute.org/publications/future-of-work-initiative/
Brookings: Lifelong Learning in an AI Economy
https://www.brookings.edu/articles/the-growing-importance-of-lifelong-learning-in-an-ai-economyWorld Economic Forum: The Reskilling Revolution
https://www.weforum.org/reports/the-reskilling-revolution-better-skills-better-jobs-better-education-for-a-billion-people-by-2030/
Worker unawareness of employer AI use
MIT Sloan / BCG: Expanding AI’s Impact With Organizational Learning
https://sloanreview.mit.edu/article/expanding-ais-impact-with-organizational-learning/