Policy paper · February 2026

AI, Jobs, and the Social Contract

01 · Summary

What the report argues

The paper argues that the effect of artificial intelligence on work is a question of governance, and that power over AI deployment is concentrated in fewer than ten firms under limited democratic oversight. It finds that disruption is spreading across white-collar, blue-collar and informal work at once: entry-level pathways are narrowing, organisational middles are thinning, and algorithmic management is changing work where jobs persist. Productivity gains remain modest and largely decoupled from wages, and women are disproportionately exposed. For labour-surplus democracies such as India, it warns that unmanaged automation risks compressing wages, weakening demand and squandering the demographic dividend. It sets out seven principles and a sequenced plan, starting with a national White Paper on AI and Work, labour observatories and mandatory AI–Labour Impact Reports for large employers.

Download the full report (PDF)
Cover of AI, Jobs, and the Social Contract

02 · Key findings

Key findings

  1. Decisions that reshape labour markets are made by fewer than ten companies, unlike earlier transitions shaped by thousands of competing firms, which the paper argues makes democratic oversight structurally necessary.
  2. Entry-level work is narrowing: Indian IT services companies have reduced entry-level roles by 20–25 percent due to automation and AI, cutting off a pathway the paper links to mentorship and upward mobility.
  3. Across advanced and emerging economies, labour-productivity growth has stayed below 1.5 percent despite heavy investment in AI infrastructure, with gains concentrated in a narrow set of frontier firms.
  4. Women hold a large share of the tasks exposed first to automation, and because their spending sustains household demand, their income loss carries macroeconomic consequences for India.
  5. NITI Aayog projects that, in a business-as-usual scenario, IT services headcount in India could fall from 75–80 lakh to 60 lakh by 2031.
  6. As immediate steps, the paper calls for a White Paper on AI and Work tabled in Parliament, labour-market observatories with quarterly public dashboards, and impact reports from large employers whose automation affects more than 1,000 workers.
Download the full report (PDF)
← All research