Why Education Systems Need Rethinking in the Age of AI. Reflections on the Asia Global Dialogue

By Bapon Fakhruddin and Joshua Xu, Asia Global Fellow

The Asia Global Dialogue in November 2026 brought together economists, university leaders and business executives to discuss growth under the theme “Growth Reimagined”. A recurring thread was whether the institutions that prepare people for work are keeping pace with a labour market that AI is changing quickly. The discussion left a clear impression that education systems need to be reconsidered more fundamentally than curriculum updates allow.

A useful starting point is the economics of asymmetric information. Michael Spence, who chairs the Advisory Board of the Asia Global Institute, received the Nobel Memorial Prize in Economic Sciences in 2001 with George Akerlof and Joseph Stiglitz for work on markets in which one party knows more than the other. In Spence’s signalling model, an employer cannot observe a job applicant’s productivity directly. A costly credential, such as a university degree, helps close that gap because more capable candidates can obtain it at lower cost.

At the Dialogue, Professor L. Rafael Reif, MIT’s President Emeritus, used this framework to make a direct point. The education signal no longer works for young people as well as it once did, and new graduates’ career paths are under severe strain. When AI tools can produce competent drafts, analyses and code, a degree conveys less about whether a graduate can do the work an employer actually needs. The information gap between employer and applicant has reopened in a new form. Employers increasingly want evidence of judgement, the ability to work with AI systems, and experience solving real problems.

A second Nobel laureate, Paul Krugman, said that the current AI build-out in the United States combines very large capital expenditure with very little job creation. Taken together with the weakening of the education signal, this suggests that the entry point to professional careers is narrowing as the value of the traditional credential declines.

The consequences could extend beyond graduates. Routine entry-level tasks have long served as the training ground through which young professionals acquire tacit knowledge, the practical understanding that is rarely written down. Reif referred to evidence that senior patent lawyers gained more from AI assistance over time than junior colleagues, because they already held the experience needed to judge the output. If firms reduce entry-level hiring, fewer people will build that experience, and the supply of capable mid-career managers and senior leaders may contract within a decade.

Spence’s opening address added an important economic dimension. He argued that the gains from AI depend less on building frontier models than on diffusion, meaning the productive adoption of AI across the whole economy. In his assessment, the global economy remains in the early dip of a J-curve, with heavy investment and limited measured productivity gains so far. Diffusion depends on people. The capacity to redesign work around new tools is formed largely through education and training. Slow adjustments in education systems are therefore likely to slow diffusion and delay the productivity gains on which future growth depends.

Several directions for education reform appear worth considering, though each requires testing in national contexts.

Credentials: If the degree is a weaker signal, education systems will need complementary ways for people to demonstrate what they can do. Verified project work, supervised practice and recognised short qualifications could provide employers with better information, provided common standards make them comparable across institutions and borders.

Place of experience in learning: Reif suggested, as a discussion question, that universities may need more hands-on experiential learning. Structured partnerships with employers, including apprenticeships and work placements built into degree programmes, would allow students to begin acquiring tacit knowledge before graduation.

Learning across a career: Mid-career and senior professionals face the same pressures as graduates, but most public education funding is concentrated on the early years of life. Policy frameworks that recognise and finance continued learning would help experienced staff adapt and would protect the leadership pipeline.

Concerns regulation and shared responsibility: Employers that automate entry-level work benefit from the experience that earlier training produced. A case can be made for rules or incentives that ask them to contribute to the formation of the next generation of professionals. Clear guidance on the use of AI in teaching and assessment is also needed so that students develop independent judgement.

For developing economies in Asia and the Pacific, the stakes are considerable. Spence noted that digital technologies are already extending financial, educational and health services to populations that traditional channels cannot reach profitably. Realising that potential will require a workforce prepared for the way AI is changing professional work. Education reform is therefore closely tied to development outcomes.