Universities Are Key Drivers in the AI Reskilling Ecosystem

As artificial intelligence reshapes global labor markets, the demand for lifelong learning and continuous reskilling has moved from a policy ideal to an economic and social imperative. Yet, across both Asia and Europe, working adults seeking to update their skills face steep structural barriers. Traditional higher education models remain predominantly front-loaded toward early adulthood, while non-traditional and disadvantaged learners (including older displaced workers, rural populations, migrants, and individuals with disabilities) are frequently excluded.

While AI holds immense potential to democratize education through adaptive learning and tailored support, technology alone will not deliver inclusive outcomes. Unchecked, the rapid rollout of AI threatens to widen historic digital divides into deep structural AI divides, where privileged institutions leverage cutting-edge proprietary models while underserved learners are left behind.

To prevent this polarization, a new white paper produced by the Asia-Europe Foundation (ASEF) argues that AI-enabled lifelong learning must be treated as an urgent policy and institutional governance challenge. Achieving equitable outcomes requires a broader collaborative ecosystem connecting higher education institutions with governments, employers, public authorities, and local communities.

Different Policy Contexts, Shared Challenges

The structural barriers blocking adult reskilling reflect varying regional governance philosophies, regulatory postures, and institutional frameworks across Europe, the UK, and Asia.

In Europe, higher education governance under the EU rights-based model is defined by strict regulatory compliance, strict data protection under GDPR, and risk-tiered alignment with the EU Artificial Intelligence Act. Initiatives like the European Union's Union of Skills and the Council Recommendation on Micro-credentials attempt to establish standardized credit frameworks. While this rights-based model offers structural protection against algorithmic discrimination, public funding structures remain heavily optimized for initial degree cycles, creating rigid accreditation loops that can slow agile reskilling provision.

By contrast, the United Kingdom relies on a sector-led model that encourages instructional flexibility and institutional self-regulation, though uneven institutional support can lead to localized fragmentation in student digital skills. Across East and South-East Asia, developmentalist frameworks, such as the ASEAN Qualifications

Reference Framework (AQRF) and the ASEAN Digital Masterplan, prioritize national competitiveness, digital literacy expansion, and developmental flexibility. Countries like Malaysia utilize university-level "regulatory sandboxes" to pilot AI integrations in real time. While this posture enables rapid curriculum adaptation, systemic resource disparities mean underfunded public institutions struggle to deploy advanced digital infrastructure equitably.

Structural Barriers

This divergence across policy environments exposes a common underlying challenge: expanding physical and digital access while mitigating systemic automated risks. Three core structural friction points impede equitable reskilling:

● Unequal access to advanced AI infrastructure: Beyond basic internet connectivity, higher education faces multi-dimensional disparities in access to advanced proprietary models, computing power, data infrastructure, and AI literacy.

● Algorithmic bias and the "black box": Predictive advising analytics, admissions screeners, and automated evaluation tools risk automating historical exclusion. Machine learning models trained on skewed administrative data systematically penalize socio-economic minorities, while AI-generated text detectors disproportionately misclassify writing produced by non-native English speakers.

● Institutional rigidity: Fixed degree structures and a lack of flexible mechanisms for the Recognition of Prior Learning (RPL) make recurring participation nearly impossible for working adults balancing full-time employment and caregiving obligations.

Institutionalizing Microcredentials

Overcoming these barriers requires complementing front-loaded degrees with flexible, stackable microcredentials. Demand-side mechanisms like Individual Learning Accounts (ILAs) grant working adults dedicated subsidies to access accredited modular training. When paired with clear regional recognition standards, short-form courses can accumulate progressively over time into formal higher education qualifications.

In parallel, institutional data governance must expand beyond technical open-data access to guarantee data sovereignty and community inclusion. By combining machine-actionable FAIR principles (Findable, Accessible, Interoperable, Reusable) with purpose-oriented CARE principles (Collective Benefit, Authority to Control, Responsibility, Ethics), public universities can ensure that educational data systems protect linguistic diversity, respect community rights over cultural assets, and serve localized learning needs rather than commercial extraction.

A 3-Step Policy Roadmap for Higher Education

To scale these institutional practices, national policymakers, university leaders, employers, and social partners across Asia and Europe must coordinate across three phased implementation periods:

1. Enable supportive policy conditions (Phase 1: 0–12 Months)

National and regional governments must establish the legislative and financial foundation for non-degree learning.

● Directly fund microcredentials: Shift public subsidies beyond full-degree programs to short, targeted AI and technical courses, utilizing individual learning accounts.

● Establish recognition standards: Build national and regional credit frameworks with clear credit values to enable seamless transfer across borders and institutions.

● Invest in open AI infrastructure: Provide low- or no-cost access to quality digital tools, Open Educational Resources (OER), and localized language models for under-resourced institutions.

2. Transform higher education provision (Phase 2: 1–3 Years)

Universities must redesign their operational models and build strong external partnerships.

● Redesign modular learning: Expand Recognition of Prior Learning (RPL) so professional experience and non-formal training count toward academic progression.

● Build cross-sector ecosystems: Partner with employers, industry leaders, public authorities, and civil society to co-design curricula that address real-time skill deficits and local community needs.

3. Sustain: Embed, monitor and improve (Phase 3: 3–5 Years)

Institutions must embed accountable governance and continuously monitor equity outcomes.

● Enforce human-in-the-loop oversight: High-impact AI decisions, such as admissions screening or automated evaluations, must require explainability, mandatory human verification, and formal appeal mechanisms for learners.

● Adopt CARE and FAIR data principles: Combine technical interoperability (FAIR) with community sovereignty, linguistic diversity, and non-extractive data practices (CARE).

● Measure equity by outcomes: Evaluate AI deployment not by speed or adoption scale, but by using disaggregated data to verify whether participation, retention, and completion gaps are actively narrowing for marginalized groups.

The Path Forward

The true value of artificial intelligence in higher education cannot be measured by processing power or administrative cost savings. It must be judged by whether it expands meaningful participation, protects fundamental rights, and drives social mobility for individuals navigating prolonged, unpredictable career trajectories.

By aligning funding mechanisms, governance safeguards, and cross-sector partnerships, universities across Asia and Europe can evolve into inclusive lifelong learning hubs, ensuring that the AI transition leaves no learner or worker behind.

N.B. This commentary draws on the ASEF white paper co-authored by Lampros Stergioulas, Cinta Gallent-Torres, Munir Moosa Sewani, Anshumaan Goel, Katerina Makri, and David Timis.