The AI Skills Gap Is Not Where You Think
Employers, students and universities are each looking in a different direction, and graduates are paying the price.
Sign up for my Substack daily AI newsletter here.
Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/
A private college chief executive wrote recently that AI skills are becoming increasingly important for graduate employability, and that fresh graduates are already launching one-person businesses with AI. The evidence supports the claim about demand. What is less obvious is where the skills gap actually sits. It is not simply that students lack AI skills. Employers, students and universities are each looking in a different direction, and graduates are paying the price.
CONTEXT AND BACKGROUND
Employers are moving fastest. In the United States, the National Association of Colleges and Employers found that more than a third of entry-level jobs now require AI skills, nearly triple the share six months earlier. Yet half of the students it surveyed said they were not building AI skills, and almost a third saw them as of little or no importance to their careers.
Meanwhile, universities sit in the middle. A global survey of more than forty-five thousand students and lecturers in thirty-five countries found that only twenty-eight per cent of students say most of their assessments reflect the skills an AI-enabled workplace needs. Only twenty-nine per cent believe their lecturers are well equipped to guide AI use.
INSIGHT AND ANALYSIS
The gap is more nuanced than it appears. GMAC’s survey of recruiters in thirty-nine countries found that employers expect AI skills to be the most valued ability in five years. Today they rank fourteenth, behind communication, problem-solving and adaptability. Fewer than half of employers believe graduates are well prepared in AI, and a third plan to replace some entry-level roles with it. Employers want both: strong human skills now and AI fluency soon.
In South Africa, the stakes are high. Graduate unemployment rose to twelve point two per cent in the first quarter of 2026, up nearly two percentage points in a year, even though a degree still sharply improves job prospects. Higher Education Minister Buti Manamela told universities this month that AI makes it more urgent to teach students to question arguments and test evidence. As he put it, AI’s “usefulness does not remove our responsibility to judge it.”
I have previously written about this in the context of how students use AI. I argued that the real risk is passive use: students who accept a model’s output learn to copy, while those who challenge it learn to reason. That distinction is exactly what employers should mean by AI skills.
IMPLICATIONS
For universities, the priority is assessment. If students are graded on work AI can produce, they will learn to produce it with AI and little else. Assessments should require students to use AI, critique its output and defend their own judgement, and lecturers need training to guide this.
Employers should be clearer about what they mean by AI skills, and should invest in graduates rather than expecting them to arrive fully formed. For students, AI fluency is no longer optional, but it is not enough on its own. Skills alone cannot solve unemployment either. As researchers at the Institute for Security Studies argue, the problem is systemic, and South Africa’s economy creates too few productive jobs.
CLOSING TAKEAWAY
The AI skills gap is real, but it is not simply a shortage in students’ heads. It sits between three groups moving at different speeds. Employers are raising expectations, students remain sceptical, and universities are still adjusting how they teach and assess. Graduates are caught in the middle. Closing the gap requires each group to move: universities to redesign, employers to invest, and students to engage. The goal is not graduates who can operate AI, but graduates who can judge it.
Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net
Comments