Future of Work in the Global South: Skills, Mobility, and Opportunity

Contents

Executive Summary

This panel discussion addresses the critical skills gap between the Global North and Global South in AI adoption, examining how the Global South lags by approximately 2x in AI adoption despite having significant demographic advantages. The panel emphasizes that AI skilling must be intentionally designed for local contexts, embedded in existing infrastructure, and reach informal workers and marginalized communities—not just elite programmers—to realize AI's potential for equitable economic mobility.

Key Takeaways

AI skills in the Global South are not an elite coding problem—they are a systems problem. The priority is embedding AI as infrastructure in tools workers already use (WhatsApp, payment apps, crop advisory systems) and reaching the 70-80% of Global South employment in MSMEs, not training the next generation of PhD researchers.
Data is destiny: Without real-time labor data tracking income outcomes in informal sectors and disaggregated by gender, governments cannot design effective policy. Start small with sector pilots, leverage existing digital rails (mobile money, payment systems), and institutionalize annual reviews rather than waiting for perfect national dashboards.
Portable, modular credentials matter more than traditional degrees because workers in the Global South move fluidly between informal and formal sectors, and between geographies. Stackable micro-credentials recognized across countries enable this mobility and resilience.
Gender is not a checkbox—it is a design variable. From the outset, ask: What framing makes this relevant to women? What tools do women in this sector already trust? How do unpaid care responsibilities shape their availability? Are we creating access through self-help groups, voice interfaces, and local languages?
By February 2027, the defining success metric should be portable AI literacy embedded in sector-specific tools reaching informal workers and women—not the number of AI bootcamp graduates or coding certificates issued.

Key Topics Covered

AI adoption disparities between Global North and Global South
Workforce skill gaps as barriers to AI adoption (40-50% of firms report inability to adopt due to lack of skills)
Gender dimensions of AI skill development and women's participation in AI-enabled work
Informal sector invisibility in labor data and policy design
Micro and small-medium enterprises (MSMEs) as primary employment in the Global South
Locally-contextualized training vs. one-size-fits-all approaches
Labor market data systems and real-time tracking of AI impact
Cross-sector partnerships (government, private sector, philanthropy, multilateral organizations)
Digital public infrastructure (DPI) approaches to AI development
Portable credentials and stackable micro-credentials for worker mobility
Voice-based AI interfaces as accessibility solutions for low-literacy populations
AI opportunity fund initiatives and scaling models across Asia-Pacific

Key Points & Insights

Skill gap is the binding constraint on AI adoption: OECD survey data shows 40% of manufacturing/finance firms and 50% of SMEs cannot adopt AI due to workforce skill deficits—this is not a fringe issue but a primary adoption barrier.
AI is not exclusively a coding problem: Only ~1% of the workforce needs AI specialist skills. The demand is for workers across all sectors who can understand data, apply AI tools to existing jobs, and recognize limitations—requiring broad-based foundational literacy rather than elite training.
The Global South has structural labor market differences requiring different interventions: High informality, significant labor mobility between sectors, and limited institutional capacity mean training must be:
- Modular and portable (credentials travel with workers, not tied to employers)
- Accessible to informal workers, platform workers, and temporary employees—not just full-time employees
- Embedded in tools workers already use daily (WhatsApp Business, digital payment apps, agricultural advisory systems) rather than standalone classroom training
Gender participation is conditional on relevance and framing: Women are not inherently less likely to participate in training than men, but they are underrepresented in STEM. However, in countries with high gender employment disparities (India, Saudi Arabia, Korea, Japan), women show higher interest in learning AI as a career advancement opportunity. Framing matters—presenting training as "how to use this tool in your job" rather than "become an AI specialist" increases participation.
Labor data systems are fundamentally misaligned with AI reality: Three critical blind spots exist:
- Informal sector invisibility: Most Global South employment is informal, yet lacks real-time productivity/adoption data
- Gender invisibility: Women's unpaid care work and home-based production remain undercounted
- Outcome distortion: Systems track enrollment/certification but rarely measure 6-12 month income mobility or employment stability
Current labor data is built for industrial-era economies, updated annually by broad sectors, while AI changes tasks quarterly.
Localized, sector-specific curricula dramatically outperform generic training: AVPN's example from Gujarat (Junior Achievement India Services training 52,000 workers) showed that a 15-hour, 2-3 day in-person curriculum with practical sector-specific outputs (e.g., creating logos, price lists, packaging designs for food processing MSMEs) and voice-based interfaces in local languages achieved relevance and adoption that broader programs missed.
AI's labor impact is nuanced and uneven: Rather than wholesale job displacement, AI is raising the bar for skills in jobs it complements (employment growth faster in AI-exposed roles). However, displaced workers will enter MSMEs and intensify competition/compress incomes—so AI skilling must be paired with productivity-focused tool adoption through fintech, agritech, and software provider partnerships.
Institutional agility is critical: Existing vocational systems (e.g., India's ITIs) are designed for slower industrial cycles, not quarterly AI changes. Annual skill gap reviews, instructor retraining every 18 months, and iterative curriculum design are necessary.
Effective design relies on intentional reach mechanisms, not assuming workers will self-select into training:
- Deliver tools through self-help groups and cooperatives (proven, community-driven structures)
- Use voice interfaces in local languages (removes literacy and language barriers)
- Track gender-disaggregated income changes (reveals where AI tools are working/not)
- Align access with microfinance networks (existing, trusted infrastructure)
Cross-sector partnership is essential but requires bridging: Government, multilateral organizations (OECD, ADB), private sector (Google, tech firms), philanthropy (AVPN), and civil society must collaborate to design interventions. No single sector can solve this alone; the panel itself exemplified this diversity.

Notable Quotes or Statements

Anjeli (CSIS): "This is not simply a skills gap. It's really a systems mismatch." — on the 54% of South Asian youth lacking skills for decent jobs.
Anjeli: "For many informal workers, AI will function as infrastructure, not as a standalone product." — on how AI adoption will actually occur in the Global South.
Chenny (Google): "AI will not replace your job but probably somebody who knows how to use AI might." — reframing the displacement concern as a competition for skills, not machines vs. humans.
Nana (AVPN): "We scale what works locally not what looks good globally." — the core principle underpinning locally-contextualized interventions.
Angelica (OECD): "Training needs to be available for temporary workers, workers in the informal sector, platform workers, not just full-time employees as it tends to be right now." — on the mismatch between current training delivery and labor market reality.
Anjeli: "You're not going to see the impact of AI on labor and on income unless you start tracking it effectively." — on the necessity of redesigned labor data systems.
Chenny: "The only thing that is constant is change." — on the necessity of agile governance structures for AI skilling.

Speakers & Organizations Mentioned

Speaker	Role/Organization
Abinit (Moderator)	Vice President, Access Partnership
Nana	CEO, AVPN (Asian Venture Philanthropy Network)
Anjeli	Senior Associate, Emerging Asia Program, CSIS (Center for Strategic and International Studies)
Chenny Yun	Public Affairs Lead, Government Affairs & Public Policy, Google Asia Pacific (Singapore)
Angelica Salvid Pero	Senior Adviser, OECD's Director for Employment, Labor & Social Affairs; leads OECD project on AI in work, innovation, productivity, and skills
Ained	Co-panelist (role/organization not fully specified)

Organizations & Initiatives Referenced:

AVPN (Asian Venture Philanthropy Network) — driving equitable development and inclusive growth across Asia since 2013
Google — issuing Google Career Certificates (1M+ globally), developing Google AI Opportunity Fund (Phase 2, $27M USD across Asia-Pacific), supporting India AI Mission, developing voice tools (e.g., health advisory tools)
Asian Development Bank (ADB) — partnering on AI Opportunity Fund
OECD — labor market research, analysis of AI adoption barriers, work on ethics of AI in the workplace
CSIS — research on technology-enabled development in emerging Asia
Access Partnership — policy research and development
American Development (USAID) — partner on AI opportunity initiatives
People's Action for Development (PAD) — local training provider delivering offline AI training in rural villages
Junior Achievement India Services — training provider in Gujarat, trained 52,000 workers with sector-specific AI curriculum
India AI Mission (launched 2024) — government initiative
Sham Shakti Niti 2025 — India labor policy framework addressing microenterprise skilling and macro labor market dynamics
Viksit India 2047 — macro mission for India's development

Technical Concepts & Resources

Tools & Products Referenced

Google Gemini — generative AI chatbot/assistant
Google Maps, Gmail, Google Photos — examples of AI already embedded in everyday products
VO3 — video creation tool mentioned as lowering barriers to entry
Healthonomy — Google AI-powered voice tool for democratizing medical information in native languages
Google Career Certificates — digital credentials integrated into India's Skill India Digital Hub
WhatsApp Business Tools — with embedded AI for customer message management
Digital payment apps — with AI-embedded features for tracking sales, inventory, pricing
Agricultural advisory systems — AI crop advisory with localized farmer guidance

Methodologies & Frameworks

AI for All Workforce Scaling Policy Toolkit — AVPN-developed guide for designing AI skilling initiatives, covering needs identification, skills gap assessment, workforce roadmap implementation, and monitoring/evaluation (QR code provided in session for download)
DPI (Digital Public Infrastructure) approach — India's model for AI development, noted as replicable globally
Labor market data twins — using AI to actively, regularly model how actual labor markets are changing
Sector-specific pilot approach — e.g., agritech platform partnerships to track adoption and yield changes
Blended quantitative/qualitative labor data — mixing statistics with granular stakeholder consultation for policy-relevant insights
Online job vacancy analysis — using AI to analyze emerging labor market needs in real-time (timely but expensive in terms of data processing and human capital)

Key Metrics & Data Elements

AI adoption gap: Global North adoption is ~2x higher than Global South
Skill barrier to AI adoption: 40% of manufacturing/finance firms, 50% of SMEs cannot adopt due to workforce skills
Gender AI technology reach: Less than 10% of AI technology in Asia reaches women (per AVPN landscape mapping)
South Asian youth skills deficit: 54% of South Asian youth lack skills for decent jobs
Informal sector employment: 70-80% of Global South employment is in MSMEs
Training effectiveness: Workers receiving employer training show better outcomes on AI impact on job quality and performance

Policy & Credential Standards

Micro-credentials — stackable, portable, employer-recognized training units
Skill India Digital Hub — India's platform integrating digital credentials
Mutually recognized regional credentials frameworks — proposed for cross-country labor mobility in Asia-Pacific
Gender-disaggregated income tracking — proposed as annual measurement standard for skilling program outcomes

Note on Transcript Quality: This transcript contains some repetition and speech artifacts (e.g., phrases repeated three times: "global south? global south? global south?"; "thing. thing. thing."). The summary reflects the substantive content while filtering out these mechanical repetitions.