AI requires universities to rethink their roles

Artificial intelligence is reshaping work, and its effects are being felt across the Association of Southeast Asian Nations and China with growing force. The China-ASEAN Expo reflects the scale of that opportunity, bringing together leaders and businesses to deepen bilateral collaboration. But ASEAN economies are adopting AI at different speeds, with China ahead, and the gap between those positioned to leverage AI and those still building the foundations is widening. For universities, that gap means ensuring graduates are equipped to lead in economies AI is transforming.

China's micro-drama industry illustrates this: In the first quarter of 2026, over 95 percent of the 128,000 micro-dramas released in China were AI-generated, prompting Hengdian World Studios to reinvent its studios into immersive tourist experiences as directors adapted rather than abandoned their craft. AI displaced the routine but human creativity and adaptability found new ways to create value. It is a pattern now repeating itself across professional services, finance, healthcare and education.

According to the International Labour Organization's 2026 report,6.1 percent of jobs held by workers aged 15 to 29 are in occupations highly exposed to AI-driven change, concentrated in administrative roles that have long been the most accessible entry points into the workforce. The Organization for Economic Co-operation and Development found in 2026 that young workers in AI-exposed roles are already bearing the brunt, with unemployment among 20 — to 30-year-olds in those roles rising roughly 3 percentage points in 2025. Roles in science, health and engineering demand deep disciplinary capability, which universities must develop.

Entry-level roles are not disappearing but evolving. Keeping curriculum content current is necessary but not sufficient. Graduates must leave with disciplinary depth, higher order skills and the enduring capabilities to think, judge and act independently.

Entry-level tasks that once served as career apprenticeships — background research, routine drafting and basic analysis — are among the first tasks AI displaces. The capacity to analyse, evaluate and reason independently has moved from what sets a graduate apart to a day-one expectation.

Beyond disciplinary theories and AI knowledge, graduates need discipline-practice knowledge: the ability to bridge theory and practice, learning from practitioners solving real problems. The Strada Institute for the Future of Work found that work experience is the single most influential factor in entry-level hiring decisions, outweighing grades and academic distinctions. What matters is not the number of internships but their quality: whether students develop analytical judgement and discipline-practice capabilities in real-world settings.

The forward deployed engineer illustrates what quality looks like in practice: a professional embedded within a client organization to help design and deploy AI in specific business contexts, bridging the gap between AI capability and real-world application. That combination of disciplinary depth, AI fluency and practical deployment experience illustrates one increasingly important pathway for graduates.

At the National University of Singapore, we are rethinking how students gain practical experience. Forward deployed settings are one avenue, where students learn alongside professionals who are actively deploying AI, rather than simply studying it.

The deeper challenge is whether universities' teaching, assessment and learning practices are keeping pace, or whether the tools being introduced are working against us. Education rests on two enduring pillars: learning and thinking. Left unchecked, AI nudges thinking towards atrophy, producing graduates who are fluent in prompting but struggle to reason without technology. This is de-skilling: capabilities that were once built through practice and effort quietly eroding through disuse.

There is a more serious risk we call never-skilling, where foundational abilities are never developed in the first place because AI performs them from the outset. A graduate who has never had to work through difficulty independently has not been deskilled by technology; they have simply never acquired the skill at all.

The OECD's 2026 Digital Education Outlook shows exactly this pattern: high school students who practised maths with AI scored 17 percent worse on closed-book exams than peers with no AI access, despite performing better on practice questions. Performance improved. Understanding did not.

Each discipline requires its own approach to embedding AI without hollowing out disciplinary thinking. Pedagogy must drive technology, not the other way around. AI literacy requires a common foundation, but deeper capability must be developed within the disciplines where students will actually apply it. Preserving the conditions for productive struggle also matters. Grappling with difficult problems independently is how genuine capability takes root.

Assessment is where that commitment is tested most directly. The goal is to use AI deliberately in teaching and learning, strengthening rather than substituting for the capabilities graduates must develop. Well-designed assessments should provide credible evidence of students' independent reasoning, sound judgement and critical thinking. For capabilities that must be demonstrated independently of AI, students can show what they genuinely know through supervised, oral, practical or other verifiable forms of assessment. Students adapt to AI far faster than assessment practices, and closing that gap is a pressing challenge for universities.

When ChatGPT burst onto the scene in 2022, this year's graduating class had just started university. They became the first cohort to complete a bachelor's degree with AI as a daily reality.

The world they are entering will keep changing, and so will AI. Learning to learn and adapt across a career spanning technological shifts is one of the most valuable capabilities a university can develop. Preparing graduates for that reality, rather than for a fixed set of roles, is the defining challenge of universities today.

Universities can rise to that challenge, but not alone. Collaboration among leading universities in ASEAN and China to strengthen AI capabilities will better position the region to leverage AI. Equipping graduates with disciplinary depth, AI fluency, the ability to keep learning and the distinctively human qualities of independent reasoning, sound judgement and ethical clarity has always been the highest purpose of a university education. In the age of AI, developing these qualities has never mattered more.

The author is the president of the National University of Singapore.

The views do not necessarily reflect those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.