What are some trends you see in AI upskilling, and education?
DD: There are a few trends that’s shaping how AVPN is thinking about the Phase 3 of the AI Opportunity Fund and each of them points to the same direction – AI literacy needs to go deeper, reach earlier, and account for who is actually being left out of the conversation.
- The shift from tool adoption to AI fluency
Optimism about AI in education is high . A recent Google research shows 77% of educators and 79% of university students believe learners will benefit from AI. Yet a massive gap exists in formal training. Students are adopting these tools, but schools are asking for the frameworks to catch up.
This is why Phase 3 of the AI Opportunity Fund moves beyond introducing workers to AI tools and focuses on building AI confidence, creativity, and critical thinking from the start of the talent pipeline. The programme positions it as a new form of foundational literacy, one that every learner, educator, and enterprise needs to participate fully in the next decade of economic and social change. The expansion is expected to benefit more than 4.7 million people in education across the region, working through educators who can embed AI fluency into everyday learning rather than treating it as a separate subject. - Governments are formalizing AI skilling infrastructure, and the question is how policy translates into delivery
Across the region, national AI programmes are scaling up. Singapore’s National AI Impact Programme, launched in March 2026, aims to support 10,000 enterprises and train 100,000 workers to be ‘AI bilingual’ – combining domain expertise with practical AI skills – by 2029. This signals that the Singapore government is moving beyond basic digital literacy toward workers who can apply AI within their own field and redesign how they work.
But well-designed national strategies face real challenges in reaching workers on the ground. The workers who most need AI skilling, especially those in small businesses, informal employment, or lower-education segments, are typically the furthest from the formal training infrastructure that national programmes are built around. Reaching them requires more than good policy design. It requires locally embedded delivery – partners with community trust, content in local languages, and training built around how people actually work rather than how a national curriculum imagines they do.
AVPN’s AI Skilling Policy Toolkit is designed to support this translation at a regional level. It is built on insights from our AI for All report and engagements with policymakers across Asia Pacific. They then inform design principles for effective AI skilling into actionable policy considerations and implementation pathways that governments can act on through the skill development cycle. The goal is to provide a common reference point that helps governments move beyond isolated pilots toward coordinated, region-wide approaches. - The MSME gap is becoming harder to ignore
96% of Asia Pacific’s companies are MSMEs, employing between 50 and 80% of the local workforce. Yet they remain systematically underserved by most AI skilling initiatives, which are designed for larger organizations with dedicated HR and training functions.
For small business owners, the barrier is not awareness of AI’s potential. It is access to training that fits how they actually operate, one that is in their language, in their sector, around the rhythms of a business with no margin for downtime. The programmes making the most progress are the ones that work through trusted local partners, deliver in local languages, and focus on practical applications specific to the sector and market. This is the design logic behind AIM ASEAN, which is operating across all ten ASEAN member states with endorsement from the ASEAN Coordinating Committee on Micro, Small and Medium Enterprises (ACCMSME), toward a target of 100,000 MSMEs. - Different workers require fundamentally different AI upskilling and education approaches
AVPN’s AI for All research identifies several distinct worker segments, each facing different barriers and requiring different programme design. It highlights why a one-size-fits-all approach to AI training does not work.
For younger workers, the challenge is depth rather than access. While 1 in 5 young adults are already participating in AI skilling programmes, engagement does not directly translate into fluency. The focus for this group needs to shift from introductory awareness toward applied, role-specific capability that translates into changed behavior at work.
For mature workers aged 50 to 65, trust and language are the primary barriers. This group is 1.6 times more likely to distrust AI and twice as likely to cite language barriers as a major challenge. 51% of workers surveyed said multilingual programmes and language support would make training more accessible. In markets like Japan and South Korea, where older workers make up a significant share of the active workforce, this is not a peripheral concern but central to any credible regional AI training strategy.
For MSME owners and workers in the informal economy, the barrier is relevance and fit. Training needs to be built around real work contexts, delivered through trusted community partners, and structured around time constraints that formal programmes rarely account for.
For workers with disabilities, the barriers are different again. Indonesia’s Annika Linden Centre and Yayasan Plan International Indonesia, both AI Opportunity Fund grantees, delivers AI training through formats specifically designed for different disability types with the goal of supporting participants into stable, digitally-enabled employment.
The premise is simple: a programme that is not designed for how someone actually experiences the world is not accessible, regardless of whether it is technically available to them.
Pages: 1 2
