Ask ten hiring managers whether new graduates need real AI skills, and most will say yes without much hesitation. Ask ten students the same question, and you'll get a lot more debate, with some genuinely unsure whether AI belongs in their work at all. That gap sits right underneath the job market at the moment. It's shrinking faster than either side really planned for, and whoever graduates into 2027 is going to feel it more than any class before them did.
Look at how quickly this shifted within a single year. Entry-level job postings requiring some AI skill went from roughly one in ten earlier in the year to well over three in ten by the middle of it, nearly tripling in a matter of months. Internship listings mentioning AI-related work have almost doubled too. This isn't a slow structural drift anyone can wait out. It's happening within a single hiring cycle, and it means the graduating class two years from now will be judged against a bar that didn't fully exist when they started their degree.
What's worth noticing is where these skills are showing up. It's not just software and engineering roles anymore. AI expectations have spread into financial services postings, marketing and media roles, coordinator and analyst positions, exactly the kinds of jobs most graduates outside computer science actually apply for. Roughly three-quarters of employers now treat AI skills as a genuine advantage or an outright requirement for at least some of their open roles, not a nice-to-have buried in the job description.
There's a part of this that gets misread constantly. None of it means a finance or marketing graduate needs to learn to build machine learning models. Job listings asking for people who can work with AI agents and automated workflows have grown dramatically over the past year. Listings looking for people who can actually engineer AI systems remain a tiny sliver of overall hiring, by comparison. What employers are really asking for is fluency. Can you get useful output from a tool? Can you tell when that output is wrong or just confidently phrased? Can you fold it into a workflow without creating more cleanup work than it saves? That's a very different skill from writing code, and it's one nearly any graduate can build regardless of major.
There's a genuine complication sitting underneath all of this, though, and it's worth being honest about. A meaningful share of employers now say AI has quietly raised the experience bar for entry-level roles, even as the basic tasks that used to teach new hires, like drafting routine documents, doing first-pass research, and formatting reports, get automated away before a graduate ever gets to practice them. That's an uncomfortable bind. Graduates are being asked to arrive more capable. At the same time, they're getting fewer low-stakes chances to build that capability on the job the way earlier cohorts did.
That bind is exactly why credentials alone won't carry graduates through 2027 the way they might have a few years back. A large majority of employers now say they rely on skills-based hiring rather than degree pedigree alone, and when asked what actually convinces them a candidate is ready, a portfolio of applied work consistently ranks alongside AI fluency itself. In practice, that means students need to walk in with evidence: a class project genuinely built using AI tools well, a writing sample that shows good judgment about what to keep and what to fix, a habit of using these tools thoughtfully rather than either avoiding them out of caution or leaning on them uncritically.
Communication and strategic thinking haven't gone anywhere on employer priority lists. In a lot of surveys, they still rank above AI fluency itself. Strategic thinking in particular grew faster in importance this year than any single technical skill did, among graduate-level hires specifically. Even MBA programmes, where AI expectations run about as high as they get, show roughly one employer in four treating AI proficiency as baseline rather than a differentiator. Good graduates in 2027 will mostly use AI to think faster, then explain what they found to someone else in plain language. That second part hasn't gotten any less important.
Give it a couple more years, and AI-ready probably stops being its own category altogether. It just folds into what employable already means, roughly how basic computer literacy did a generation back, not something worth its own bullet point on a resume anymore. Treat it as baseline starting now, rather than as an extra, and the jump from campus to a first job should feel less rough than it did for the class right before.
(The author is CEO, SHRM APAC & MENA)
The opinions expressed in this article are those of the author and do not purport to reflect the opinions or views of THE WEEK.
