By Emil Bjerg, journalist and editor
Your teenager is staring into a decision that used to feel straightforward: university, degree, career – a tried and tested, relatively linear path to success. But more and more new graduates struggle to find footing in the labour market, and AI appears to be the culprit. So – should your child still go to university in the age of AI? We’ve got you covered.
Something is happening in graduate employment, and economists and labour market researchers across the world are paying attention to it. In the United States, the Federal Reserve Bank of New York found that the unemployment rate for recent college graduates aged 22 to 27 ended 2025 at 5.6% – well above the average rate of 4.2%, and the highest December figure for recent graduates since 2020.
The picture in Asia is starker still. In China, youth unemployment peaked at 18.9% in August 2025 – a record high – as the country produced 12.22 million university graduates in a single year, the largest cohort in its history. Job postings for graduates fell by 22% in the first half of 2025, while the number of applicants rose by 8%. The mismatch has driven a significant number of graduates into delivery and gig work – on China’s two largest platforms, Meituan and Ele.me, more than 20% of riders hold a university degree: a dramatic collapse in graduate expectations from a generation that enrolled in higher education expecting to enter the knowledge economy.
In Europe, the German Institute for Employment Research projected in 2025 that 1.6 million jobs could be reshaped or lost to AI in Germany alone over the next fifteen years.
The New Normal for Entry-Level Jobs
The learning curve between entry-level and experienced workers – the traditional on-ramp that graduates have always been able to rely on – is being challenged. AI is automating the mundane tasks that used to teach people the basics of their profession. That means the graduates who succeed will be those who arrive already able to think, ideate and analyze – not just execute.
IBM announced in February 2026 that it would triple its entry-level hiring – but only after rewriting every entry-level job description from scratch. Nickle LaMoreaux, IBM’s chief human resources officer, put it directly: “The entry-level jobs that you had two to three years ago, AI can do most of them. So, if you’re going to convince your business leaders that you need to make this investment, then you need to be able to show the real value these individuals can bring now.”
The new graduate roles focus less on routine coding and administrative tasks, and more on customer interaction, AI oversight, and the kind of contextual judgement that machines cannot provide. Dropbox is expanding its internship and graduate programmes by 25% for a related reason: its chief people officer observes that younger workers’ AI fluency gives them a genuine advantage over more senior colleagues. In other words: graduates are not unwanted – but a new type of preparation is needed.
Joseph Fuller of Harvard Business School and Matt Sigelman of the Burning Glass Institute, writing in the Harvard Business Review, put it precisely: generative AI will affect some 50 million jobs, but its biggest impacts will be less on the number of jobs than on the level of expertise required to do them.
And that’s actually good news for students and new graduates. According to Melanie Rosenwasser, chief people officer at Dropbox, young people are actually entering the workforce with better AI skills than their older colleagues: “It’s like they’re biking in the Tour de France and the rest of us still have training wheels. Honestly, that’s how much they’re lapping us in proficiency.”
How Students and Graduates Can Adapt
The evidence points toward three interconnected things that students, and perhaps their worried parents, need to think about – and none of them is the dated advice of simply “study computer science.”
1) Build AI fluency early, and keep building it
The World Economic Forum’s Future of Jobs Report 2025 projects 78 million new roles by 2030 alongside significant structural disruption, and emphasises that the gap between those who can work effectively with AI and those who cannot is widening rapidly. Rather than meaning “knowing code”, fluency means the ability to use AI tools to augment research, drafting, analysis, and problem-solving – and sometimes knowing when not to. The Cengage 2025 Graduate Employability Report found that only 51% of recent graduates believed they had sufficient AI skills for the jobs they applied to. That gap presents an opportunity for students and graduates. Students who close it during their degree rather than after graduation will enter a materially different labour market.
Enrolments in generative AI courses reached 15 per minute in 2025, up from eight per minute in 2024, according to Coursera. More tellingly, enrolments in critical thinking courses among people in data-related roles rose by 168% year on year. The students who are adapting most effectively are building both technical and human skills simultaneously.
2) Prioritise the skills that AI cannot replicate
The UK government’s own skills projections, published in January 2026, identify non-cognitive skills – empathy, creativity, critical thinking, interpersonal communication – as the most durable assets in an AI-reshaped economy. The OECD Skills Outlook 2025 makes a similar point: as AI automates routine tasks, demand shifts toward higher-order cognitive abilities. The WEF Future of Jobs Report reinforces this, ranking analytical thinking as the single most important core competency for employers globally – above AI and big data skills individually.
These are not skills that emerge from an online course or a bootcamp. They are developed through years of sustained intellectual and social engagement – through arguing a thesis, confronting evidence that contradicts a hypothesis, and learning to reason under uncertainty. And a lot of it is learned through years of socialising in high school and college.
Research published by Gerlich in 2025 adds a counterintuitive finding: higher education acts as a protective buffer against cognitive offloading – the tendency to delegate thinking to an AI tool at the expense of developing independent reasoning. Younger individuals without degree-level education showed the highest rates of AI dependence. The student who learns to use AI inside a demanding academic environment may be better positioned than the one who uses it to avoid demanding academic environments entirely.
3) Choose course and institution with real specificity
The blanket credential is weakening. UK government data from the Department for Education shows that graduates still significantly outperform non-graduates – an unemployment rate of 5.5% against 8.1%, and median salaries of £42,000 against £30,500. The degree has not collapsed in value. But it has become more variable. A degree from an institution that embeds AI literacy across its curriculum, maintains strong employer relationships, and requires students to engage with real problems during their studies is worth considerably more than one that does not. Parents should ask institutions direct questions: how is AI integrated into teaching? What does a graduate’s first year of employment actually look like? The answer will tell you more than any ranking.
Demand for formal degrees is falling fastest for AI-augmented roles, according to the PwC Barometer – down 7 to 9 percentage points since 2019. Employers are increasingly asking for demonstrated competence over credentialled completion. That does not mean skipping university. It means using it differently: building a portfolio of real work, not just a transcript.
Economic Outlooks for AI Jobs
PwC’s 2025 Global AI Jobs Barometer, based on an analysis of nearly a billion job advertisements across six continents, contains a finding that every parent and student should understand. Workers with AI skills command a 56% wage premium over peers in identical roles who lack them – up from 25% the previous year. Jobs requiring AI skills are growing 7.5% year on year, even as total job postings fell by 11.3%. And the skills required in AI-exposed roles are changing 66% faster than in less-exposed ones.
This is not a uniquely Western or Eastern finding. The PwC Barometer draws on data from six continents, and the wage premium for AI skills holds across every industry analysed – from financial services in Hong Kong to manufacturing in Germany. The competitive advantage of AI fluency is a global shift in what advanced economies reward.
The Long View
Jensen Huang, the founder and CEO of Nvidia – the company whose chips underpin virtually every major AI system in the world – put it with characteristic directness: “You’re not going to lose your job to an AI, but you’re going to lose your job to someone who uses AI.” This is not a reassurance from someone with an interest in downplaying disruption. It is an assessment from one of its engineers.
The Lumina Foundation, one of the most rigorous research bodies on higher education, has consistently argued that the narrative around degrees becoming obsolete misrepresents the available evidence. Rather than considering not going to university, bright students should consider which university to attend, what to study, and what to build alongside it.
The students who will benefit from AI, rather than being stopped by it, are the ones who used their years in higher education to develop the capacity to think clearly, adapt quickly, and work alongside powerful tools – without being diminished by them. That combination – human depth and AI fluency – is, on the available evidence, rarer and more valuable than it has ever been.
