As India’s workforce in low-skill and low-value-added services remains vulnerable to AI, robust enabling institutions are essential to help transition workers to medium- and high-skilled jobs, where AI can augment their efforts rather than replace them.
“Insuring Institutions are intended to provide a soft-landing for workers whose finances have been hit and whose well-being has been affected during the transitionary period. These institutions help secure a standard of living during the shift, keep inequalities in check and aid in keeping the social fabric cohesive,” Economic Survey 2025 said.
They are also responsible for reducing the recession risk, for individuals and for societies.
Insuring Institutions build safety nets (such as the National Insurance Act, 1911 and the Beveridge Report, 1942 in the UK), protect worker rights, provide finances (such as credit unions), housing (the emergence of the Young Men’s Christian Association during the 19th century), and social and emotional support during periods of displacement.
“We propose the necessity of a third institution, Stewarding Institutions… It is important to clarify that stewarding in this case does not imply placing restrictions on innovation or dictating a narrow set of applications for technology. It implies that policymakers demonstrate a certain degree of cognisance when it comes to emerging technologies, so that when the need arises, they stand well-placed to mitigate any adverse effects that emerge as by-products of technological applications,” the Survey said.
These institutions would be agile, crosscutting across sectors and up to date on the latest developments, so that they are equipped to identify both opportunities and threats. Stewarding institutions will have to be responsible for designing an approach that delicately balances public welfare without stifling innovation.
They are also required for fostering the social acceptability of AI by promoting the right levels of transparency and accountability in AI applications. For instance, AI applications in healthcare, health insurance and education would greatly benefit from higher-degrees of transparency and accountability, as these sectors are way more human-centric relative to others. Biases in models being applied to these sectors can result in adverse, unintended consequences.Labour in the AI era
Building each of these institutions is a time-consuming process due to the challenges involved in mobilising the required intellectual and financial resources. the Survey said, adding that their need is not immediately clear and aligning our social structures, regulators, cooperatives, and policymaking institutions towards a distant goal is often an uphill battle.
However, unlike previous technological revolutions, advancements in AI today and the utility it has the potential to deliver are being shared by all countries thanks to a globalised world. AI presently being in its infancy has made it so that every nation, irrespective of its income level is on equal footing as far as discovering its applicability goes, it said. This level playing field provides India the time to build the necessary institutions that will minimise the disruptions and maximise societal benefits.
“Another reason we postulate that India has the time available to reinforce and build supporting institutions is due to the nature of technological revolutions. For technologies to be widely adopted, they have to overcome several obstacles that make them ubiquitous. Historically, technologies that have positioned themselves as General Purpose Technologies have done so through refinement and increasing cost-effectiveness. This then created the demand for the technology, reorienting macroeconomic processes around it by changing the way we invest, what we demand, the education we impact and the quantity of labour we demand,” it said.
The road to a reorientation of this size is long and we hypothesise that it could be no different for scaling AI adoption. For AI to be widely applicable in the real world, it needs to be made practical, reliable, requires the establishment of supporting infrastructure and has to achieve sufficient levels of resource efficiencies to be viable, it said.