India AI Impact Summit Insights
All sessions

AI, Labor & Inclusive Growth: Policy Pathways for the Global South | India AI Impact Summit 2026

Contents

Executive Summary

This panel discussion explores how countries in the Global South can harness AI to create inclusive economic growth while protecting vulnerable populations from labor displacement and wealth concentration. Speakers present concrete case studies—from Togo's AI-powered cash transfer system to India's digital public infrastructure—and argue that bottom-up, multistakeholder approaches centered on worker agency and human trust are essential to ensuring AI benefits reach marginalized communities rather than concentrating wealth.

Key Takeaways

  1. AI can be a powerful tool for identifying and reaching the poorest if designed with appropriate data sources (satellite imagery, consumption patterns) and paired with trusted human intermediaries—but must be monitored to ensure it outperforms existing systems.

  2. The net employment impact of AI in the Global South depends on policy choices: governments must actively invest in curriculum reform, talent attraction, reskilling programs, compute infrastructure, and data access to ensure developing economies capture new job creation rather than only experiencing displacement.

  3. Cash transfers—especially large lump-sum transfers—should be a cornerstone of AI-era social policy because they empower recipients with agency to adapt to economic change, are evidence-backed, and are implementable at scale in countries with modern payment infrastructure like India's UPI.

  4. Policymakers must insist on bottom-up, multistakeholder design processes that include workers, civil society, and affected communities from the start, not as afterthoughts—technology designed in isolation from ground-level realities will embed biases and fail the marginalized.

  5. Successful transitions to new economic paradigms require political will to include and integrate new constituencies—historical precedent (Switzerland's inclusion of industrial workers) and contemporary evidence (labor organizing, cash transfer responsiveness) show that fairness and dignity in transitions reduce resistance and improve outcomes.

Key Topics Covered

  • AI-driven social protection systems: Using machine learning and satellite imagery for poverty targeting and cash transfer programs
  • Digital public infrastructure: Building payment systems and data architectures that serve rural and underserved populations
  • Labor market disruption & job creation: Net job creation potential of AI (170M new jobs vs. 60-70M displaced) and new role categories emerging in the AI economy
  • Cash transfers as redistribution & resilience: Evidence for unconditional and conditional cash transfers in enabling economic adaptation and building resilience
  • Skills gap and reskilling: Challenges in preparing existing and new workforces for AI-era employment in the Global South
  • Multistakeholder governance: Bringing together governments, civil society, workers, employers, and tech companies to co-design AI policies
  • Cultural approaches to economic transitions: Lessons from Switzerland on cooperation, consensus-building, and bottom-up problem-solving during technological change
  • Trust, human intermediaries, and dignity: Ensuring technology solutions don't displace the human relationships and trust necessary for effective service delivery

Key Points & Insights

  1. AI can identify poverty more accurately than outdated government registries: Togo's machine learning system using satellite imagery and telecom consumption data identified the poorest populations more effectively than existing social files, enabling more efficient targeting of the 920,000 beneficiaries (25% of Togo's population) reached through digital cash transfers.

  2. New AI job categories are emerging at scale today: Enterprise AI roles (AI product managers, prompt engineers, AI ops engineers), frontier AI roles (at intersections of AI with cybersecurity, quantum, haptic tech), and AI-for-AI research roles are being created faster than traditional roles are being displaced; frontier AI job openings doubled in just 7-8 months during the study period.

  3. The Global South faces a "stock and flow" problem in AI workforce development: Existing tech workers lack training to transition to AI roles, and new college graduates aren't being equipped with AI skills because curriculum integration is slow, there are insufficient AI PhDs, and talented researchers migrate to Western countries due to better incentives, compute infrastructure, and datasets.

  4. Trust and human intermediaries remain critical in digital systems: Despite technological sophistication, the Digital Empowerment Foundation's success in reaching 3,000 rural locations depends on local intermediaries who build trust—technology enables but cannot replace human relationships, especially for vulnerable populations.

  5. Cash transfers are the most evidence-backed poverty alleviation tool, with new transformative potential: Large lump-sum cash transfers, demonstrated especially since COVID-19, enable recipients to make productive investments, start businesses, migrate, and adapt in locally appropriate ways; Give Directly has delivered ~$1 billion (₹10,000 crore) globally with measurable resilience and entrepreneurship outcomes.

  6. The framing should shift from "future of work" to "future of income": Rather than asking how AI will displace jobs, policymakers should focus on establishing a minimum income floor so vulnerable populations are protected regardless of which sectors or workers AI affects, and have agency in how they respond to economic transitions.

  7. Bottom-up design and multistakeholder collaboration are essential: Partnership on AI's work shows that bringing together competing interests (tech companies, civil society, labor organizations, governments) in uncomfortable conversations yields shared prosperity guidelines and job impact assessments that center worker well-being alongside productivity.

  8. Global South countries need AI talent attraction and retention strategies: China's recent talent visa program successfully repatriated diaspora researchers; India and other developing economies need comparable incentive structures, compute infrastructure, and open-source datasets to compete for frontier AI talent and avoid brain drain.

  9. Economic transitions require political coalitions of the willing: Switzerland's historical experience (labor organizing during industrial revolution, subsequent political inclusion of workers) demonstrates that societies adapt successfully to technological disruption when workers organize, demand fair shares of productivity gains, and political systems accept and integrate new constituencies.

  10. Sovereignty and policy learning require international collaboration, not imposition: Successful policies like Togo's digital cash transfer and India's UPI should be shared and adapted by other countries; one-size-fits-all solutions pushed from Silicon Valley or Western capitals will fail to account for local contexts and may embed bias.

Notable Quotes or Statements

  • On AI-driven poverty targeting (Togo speaker): "The people selected through the machine learning were really the poorest of the people... compared to the ones that were initially part of the file, which shows how effective that program has been."

  • On the primacy of human trust in digital systems (Digital Empowerment Foundation speaker): "Howsoever digital you can make it, trust will come at the end of it. You need a human loop of trust. That's the most important part."

  • On AI benefiting the Global South (BCG speaker): "The question for us is not about how AI will let go of jobs. It's more about how the global south gets its fair share among the opportunities of jobs that AI will create."

  • On shifting the narrative (Give Directly speaker): "There is a lot of talk about what is the future of work in the AI era... We think that the emphasis should be more on what is the future of income."

  • On designing with workers, not for them (Partnership on AI speaker): "Make sure these multistakeholder processes are designed with workers not for them but with them engaged at the beginning so their voices are heard."

  • On why cooperation matters (Switzerland speaker): "If you want to be free, if you want to determine your life, you need to cooperate with the people around you... You can't wait for somebody in the capital to come and help you."

  • On digital public infrastructure (Digital Empowerment Foundation speaker): "If AI is getting into becoming part of the big game of public infrastructure, the digital content must be reliable, must be ethical, must be serving, must not be capitalist, must not be biased."

Speakers & Organizations Mentioned

  • Stella Biderman – Give Directly (global cash transfer organization, ~$1B/₹10,000 crore delivered)
  • Digital Empowerment Foundation – Operating in ~3,000 rural locations across India
  • Togo government representative (speaker partially anonymized in transcript) – Designed and implemented digital cash transfer program reaching 920,000 people (25% of population)
  • Rebecca Henderson – Partnership on AI (140 organizations globally across 18 countries; leads future-of-work initiative)
  • Thomas Minder – Switzerland (perspective on economic transitions and multistakeholder cooperation)
  • Asana (or similar name in transcript) – Digital Empowerment Foundation speaker
  • BCG/MIT speaker (identity not fully clear) – Authored job creation in AI economy report
  • University of Berkeley – Collaborated on social policy lab and machine learning experiments with Togo
  • Give Directly – Mentioned as world's first and largest organization dedicated to cash transfers
  • World Bank – Funded Togo's social resilience and social protection information system
  • World Economic Forum – Referenced report on job disruption (60-70M) vs. creation (170M)
  • Esther Duflo (implied reference, though transcript says "Estu Prize Nobel") – Economics Nobel laureate at UC Berkeley, consulted on cash transfer policy design

Technical Concepts & Resources

  • Machine learning algorithms for poverty mapping: Satellite imagery analysis + telecom consumption data (call detail records) to rank districts by poverty and identify individuals living below $1/day
  • Biometric ID systems: Use of voter/citizen biometric IDs as foundation for digital identity for cash transfer programs
  • USSD (Unstructured Supplementary Service Data): Communication protocol used to contact identified beneficiaries for program enrollment
  • UPI (Unified Payments Interface): India's real-time payment system; referenced as inspirational model for payment infrastructure in developing economies
  • Dynamic social registries: Data systems that segment and update population categories in real time to enable adaptive, shock-responsive targeting
  • Job Impact Assessments: Tool developed by Partnership on AI to evaluate potential impacts on job categories before AI deployment
  • Shared Prosperity Guidelines – Partnership on AI framework for centering worker well-being in AI deployment
  • Satellite imagery for poverty mapping – Technology used in Togo case study
  • Telecom CDR (Call Detail Records) analysis – Consumer consumption pattern inference from mobile phone usage data
  • Data labs – Institutional structures (e.g., partnership with UC Berkeley) for implementing AI/ML experiments in real-world policy contexts

Reports & Frameworks Referenced

  • BCG/MIT report on "Job Creation in the AI Economy" (available for download)
  • Partnership on AI's "Shared Prosperity Guidelines" (online and publicly available)
  • World Economic Forum report on job disruption/creation figures
  • "Social protection information systems" design documentation (Togo case)

End of Summary