A few weeks back, two influential voices in India’s economic landscape, N Chandrasekaran, Chairman of Tata Sons, and Chief Economic Advisor (CEA) V Anantha Nageswaran, offered important and complementary perspectives on artificial intelligence (AI).AdvertisementChandrasekaran, in the TCS’ FY2025 annual report, described Generative AI (GenAI) as a “civilisational shift,” with the rise of AI agents and autonomous robots ushering in a future of “dark factories” and AI-assisted enterprise functions. He also highlighted the “human+AI model” of delivering solutions. Meanwhile, CEA Nageswaran, speaking at the CII Annual Business Summit, issued a note of caution: AI deployment is not inevitable. He reminded the private sector that India needs to create at least 8 million jobs a year. Hence, businesses must consider where to stop automating and instead choose labour.This tension, between AI’s promise of productivity and its peril for employment, is now central to India’s growth trajectory. The Indian Economic Survey 2024–25 had flagged the impact of AI on labour markets as a policy imperative. As AI becomes more capable and less costly, low-value service jobs, especially in India’s labour-surplus economy, become increasingly vulnerable. An IIM Ahmedabad 2024 survey found that 68 per cent of Indian white-collar workers expect their roles to be partially or fully automated within five years; 40 per cent believe AI will make their skills redundant.Clearly, India must act. But rather than resisting the AI wave, the real challenge lies in shaping it so that technology augments, rather than replaces, human potential.AdvertisementThree pillarsIndia needs a three-pronged institutional architecture, one that enables workers through education and skilling, insures against displacement, and stewards the broader social and economic transition.The enabling agenda is critical. In Learning by Doing, the economist James Bessen shows that technologies like the spinning mule took several decades to be fully adopted during the Industrial Revolution, not due to access issues but because it took time for workers and firms to learn how to use them effectively. What about AI? Unlike the spinning mule, AI is not a single tool. It is a fast-growing mix of powerful and diverse technologies. This increases both the challenge and the opportunity.Also Read | The right way to study AIIndia’s skilling efforts must be agile and keep pace with this growth in AI. India should ensure that the benefits from AI do not accrue just to a narrow band of skilled workers. Instead, vocational training, on-the-job learning, and open knowledge-sharing on leveraging AI on the factory floor and across all services must be embedded into national skilling programs.At the same time, the government must insure against job losses and dislocation. Workers affected by automation need social protection, access to reskilling pathways, and incentives to transition to adjacent roles. The skilling effort cannot be limited to college education; it must also accommodate informal workers and mid-career transitions.India also needs stewarding institutions that ensure AI is deployed responsibly, transparently, and inclusively. They are tasked with identifying emerging risks, conducting foundational safety research, and setting standards. They should also research how to build effective human-AI teams, based on a deep understanding of socio-economic value, the availability of human skills, and evolving capabilities of AI. This will be essential for designing job roles and workplace structures that make the most of co-intelligence.From automation to augmentation via co-IntelligenceAugmentation means AI systems assist humans, enhancing their judgement, creativity, and productivity, rather than replacing them outright. Take agriculture. Instead of replacing existing workers, AI-based agri-chatbots can empower farmers with timely advice on weather, pests, and crop management.In education, AI can help teachers identify student needs and personalise lesson plans. A study by Anthropic found that 57 per cent of tasks completed by Claude.ai involved human-AI collaboration, not substitution.What can enterprises do to achieve this human amplification with AI? In The Co-Intelligence Revolution, Venkat Ramaswamy and Krishnan Narayanan suggest that every organisation must: Become co-intelligent enterprises, where value is co-created between humans and AI; Reimagine their workers not as passive operators of systems but as creative experiencers (individuals who actively shape and are shaped by their interactions with intelligent technologies); Create a Co-Intelligence Knowledge Environment, where human insights, experiential feedback, and AI-driven suggestions flow dynamically to inform decisions and design.Siemens exemplifies this shift through its industrial metaverse, where engineers, designers, and shopfloor workers collaboratively engage with digital twins and AI co-pilots in a virtual simulation environment. In one compelling instance, a new factory was built entirely in the metaverse before physical construction began. Workers explored the virtual factory, offered feedback on ergonomics, workflows, and safety, and their suggestions were integrated into the final design. The actual factory space, thus, reflected their lived experiences and needs.This approach not only optimised operations but also fostered a sense of ownership, dignity, and well-being among the workers – hallmarks of a truly co-intelligent enterprise. Indian businesses should thoughtfully design co-intelligence into their environments.most readBut markets don’t always favour augmentation. Economists Daron Acemoglu and Pascual Restrepo argue that when automation becomes the dominant paradigm, innovation and investment naturally follow it, even when augmentation via co-intelligence may yield higher social benefits. One of the most important policy nudges for the Indian government would be to steer the AI solutions towards augmentation, through public-private partnerships, incentives for augmentation-based innovation, and “human-in-the-loop-of-AI-systems” design mandates. This holds especially true for contexts where automation may have a high social impact.India can and must shape the trajectory of this emerging general-purpose technology and push the AI ecosystem in a more inclusive direction.Balaraman Ravindran is Head, Wadhwani School of Data Science and AI & Centre for Responsible AI, IIT Madras. Omir Kumar is Policy Analyst, Centre for Responsible AI, IIT Madras. Krishnan Narayanan is Co-founder and President of itihaasa Research and Digital