India should follow AI-for-Labour Model for job growth, says ICRIER policy brief

Even as the adoption of artificial intelligence (AI) is stirring up debates and anxiety over possible job losses across sectors, Delhi based think tank ICRIER’s Prosus Centre for Internet and Digital Economy (IPCIDE) proposes an AI-for-Labour Model which keeps productivity (P), inclusivity (I) and entrepreneurship (E) – PIE - as core objectives of India’s AI led change.

A policy brief published by IPCIDE early this month suggests that India should embed a PIE-AI framework into its industrial, innovation, and labour policy to convert AI’s efficiency gains into broad-based growth and employment resilience.

First of all, AI adoption must be steered towards augmentation rather than substitution, it says. “Policy choices should prioritise AI applications that complement human labour, enhance task quality and productivity, and support decision making. While some roles will be automated, firms should be encouraged to create new and complementary tasks to absorb displaced workers. Targeted incentives and deliberate support from incubators, accelerators, and mentorship networks should also be provided for creating new opportunities through entrepreneurship”, the policy brief prepared by IPCIDE Professor Payal Malik and IPCIDE Fellow Nikita Jain, suggests.

Second, the researchers want digital public infrastructure (DPI) to be leveraged to widen diffusion and prevent concentration. Stating that large firms are better positioned to use AI, they point out that without intervention, productivity gains will remain concentrated to the big players. “Expanding access to basic AI infrastructure, such as shared computing facilities, cloud credits, interoperable digital platforms, and relevant data, is essential for enabling MSMEs and startups to participate in AI-enabled growth. DPI should be viewed as an enabler that supports AI diffusion, inclusion, and entrepreneurship”, the policy brief says.

The other suggestion is to see AI ecosystems are shaped to contextual, interoperable, and competitive ecosystems to democratise AI resources. “This requires advancing an AI-for-Bharat strategy that combines vernacular and domain-specific AI models with pro-competition safeguards. Supporting Indian small language models and open, interoperable AI infrastructure can lower entry barriers and enable MSMEs and startups to innovate. At the same time, competition frameworks for foundational AI and cloud markets are necessary to prevent concentration and to ensure that AI-led productivity gains are widely diffused”, the policy brief suggests.

It also notes that labour-market adjustment requires large scale, adaptive skilling rather than one-time reskilling. “Task-based models suggest that some displacement can be offset by creating new, non-automated tasks, but only when technologies are consciously designed around human–machine complementarities. For India, this implies sustained investment in vocational training, RPL, apprenticeships, and modular skilling pathways that allow workers to move across industries, sectors, and roles”, the report says.

Finally, it suggests that for AI to be adopted in a way that supports jobs, adequate infrastructure (reliable power, high-speed internet, and access to compute resources) is required. Recent research on AI readiness underscores that nations with robust digital infrastructure and extensive skill sets are better equipped to convert AI exposure into productivity enhancements while mitigating labour market shifts, the policy brief says.