The hollow hype over India as the ‘AI use case capital of the world'

Scroll.in

28-05-2025

In February 2023, Indian Prime Minister Narendra Modi called on citizens to 'identify 10 problems of the society that can be solved by AI'. In 2024, Nandan Nilekani, the IT czar who has been a driving force behind India's digital journey over the past 15 years, declared that India would soon become the ' AI use case capital of the world '.
In January 2025, the Ministry of Electronics and Information Technology's IndiaAI Mission issued a call for proposals to build Indian foundational models, the software that underlies contemporary generative AI development. One of the criteria was 'identifying and elaborating use cases that address societal challenges at scale.' And last month, the Gates Foundation and the IndiaAI Mission announced a partnership on 'AI solutions for better crops, stronger healthcare, smarter education & climate resilience'.
The discourse of AI use cases for socio-economic development is one of the most distinctive features of India's AI policy. Its promise is that AI will 'solve' difficult problems in classic sites of postcolonial development: agriculture, health, and education. It discursively links socio-economic development in rural India to industrial strategy at the cutting edge of global AI technology.
However, lacking an account of political economy, the 'use cases' approach makes for a poor policy programme. Instead, this seductive vision serves as a hype machine, paying lip service to development to legitimise a range of other interventions, from claims to geopolitical leadership to the marketisation of populations new to the internet.
The discourse of AI use cases has been foreshadowed by the digitalisation of development that followed the Aadhaar digital identity system in India. Launched in 2009 as an intervention that promised to streamline India's rights-based welfare apparatus, Aadhaar brought hundreds of millions of Indians into the purview of digital systems. Its promoters in the software industry used the promise of financial inclusion , especially following the founding of the IndiaStack project in 2015, to legitimise granting the Indian software and financial industry access to these digitalised Indians as customers. With access to private credit, impoverished Indians would now, in the financial inclusion playbook, ' enterprise themselves out of poverty '.
While poverty remains an enormous challenge , state investment in public-private infrastructure has undergirded an expansion in the software industry, spawning, for example, a new fintech industry. The targeting enabled by Aadhaar and similar systems also heralded the rise of the BJP's ' new welfarism ', shifting away from public goods like public health and primary education to the provision of cash transfers for private goods like gas cylinders. Rebranded as ' Digital Public Infrastructure ', these systems are being exported around the world as a model for the use of technology in development. Building on this approach – and promoted by a similar set of actors in the state and industry – the AI 'use case' discourse frames India's societal challenges as a resource for software capitalists.
In practice, AI use cases in these domains are largely speculative. In agriculture, for example, dozens of vernacular language chatbots promise to better inform farmers about weather conditions and planting times. In healthcare and education, the promises of AI are largely in streamlining administrative processes, such as hastening India's transition to digital health records, which is supposed to improve efficiency while delivering huge amounts of data to hospitals and insurance providers.
Despite the lack of tested applications, the discourse of AI use cases portrays the numerically vast market constituted by the poor as a national opportunity in the global AI arms race. The AI supply chain – composed of datasets, models, and computing power – is controlled by just a handful of US Big Tech actors, eliciting industrial policy responses from several states . In India, the poverty market – where poor people are figured both as users and providers of data – is imagined as a driver of growth that can give the nation a competitive advantage in an increasingly concentrated global AI market. To be sure, Indian AI industrial policy also relies on more traditional tools, including massive incentives to build domestic capabilities in semiconductor manufacturing, cloud resources, and models.
But the national champions of the Indian AI economy are imagined in the 'use case' discourse as emerging from software applications for socio-economic development. 'To [unlock] India's potential with AI', Nilekani proclaimed in 2023 , 'the trick is not to look too hard at the technology but to look at the problems people face that existing technology has been unable to solve'.
The promise of use cases, in other words, blurs the lines between the marketisation of poverty and national industrial policy that hopes to make India globally competitive in cutting edge technology.
Of course, this is too good to be true. The discourse of use cases ignores the political economy of both the AI industry and of development.
From the perspective of AI sovereignty – a major focus in India's technological doctrine – it will do little good to become the 'use case capital of the world' if semiconductors, cloud resources, and models remain concentrated in the hands of US Big Tech. Today's generative AI, even more than other digital technologies, runs on semiconductors sold by a single company – Nvidia – which are fabricated by a single factory in Taiwan – TSMC – on equipment made by one Dutch manufacturer – ASML. Meanwhile the cloud computing data centers and models required by AI are overwhelmingly controlled by Amazon, Microsoft, and Google.
AI use cases are imagined as a way to promote the growth of the domestic startup industry, but most Indian startups in the AI space and beyond don't appear to be interested in the poverty market. A 2024 survey of over 120 generative AI startups in India, which have collectively raised over $1.2 billion in the last five years, showed that 70% are providing solutions only for enterprise clients. In keeping with Indian tech's historical bias toward enterprise services, the industry appears to be largely focused on backend software components for use in industry, not consumer-facing software products, let alone for socio-economic development use cases.
This is reflected in the sectoral data. Despite the buzz, agriculture does not figure as one of the top five sectoral applications for generative AI startups. While education and healthcare do figure in the top five, these are lucrative markets for the middle and upper classes; it appears unlikely (though we need further data to definitively conclude) that AI startups in these sectors have developmental goals. This makes financial sense for startups and venture capitalists. The Indians who would be targeted by the proclaimed AI use cases are, after all, very poor with little spending power to sustain startup business models. As a recent venture capitalist report put it , the poorest billion Indians are 'unmonetisable' for startups.
AI use cases are also the wrong answer to issues of socio-economic development. Entrenched developmental problems in agriculture, health, and education need structural reforms rather than the quick technical fixes promised by AI . Indeed, the evidence over the past decade shows that reliance on digital systems such as Aadhaar to solve developmental challenges may have harmed the poor more than it helped them . Perhaps most of all, after decades of economic growth concentrated in low-employment sectors like software, Indians need mass employment, which AI use cases will not provide.
We should understand the focus on use cases, then, as a particularly Indian species of technology hype, an inflated promise that makes things happen. In the US, AI hype has most often been premised on the emergence of an 'Artificial General Intelligence' with unimaginable, humanity-threatening capabilities that is supposedly right around the corner. These inflated promises have driven a massive surge of speculative investment and pushed market valuations of AI companies to new highs, despite little proven demand for the technology.
In contrast, development as AI hype appears to offer a reasonable and socio-economically grounded alternative. Oriented not only toward the future but also toward the periphery of the capitalist system, it promises that those who have been on the margins of economic growth can serve as a source for data and a market for AI applications.
This discursive structure may not be driving massive investment similar to US AI hype. Nevertheless, it serves a range of powerful constituents:
1. For the ruling government domestically, it projects an image of benevolent, technocratic developmentalism. Alongside its Hindu nationalism, this high-tech image has been a key plank of the current government's appeal . It is no accident that the exemplary AI use cases are chatbots, which are personalised technologies that provide a one-on-one interface with citizens to access targeted services. As such, AI use cases track with the BJP's shift away from the provision of public goods like basic health and primary health to the techno-patrimonial provision of private goods under Modi.
2. Globally, the 'use case' hype enables India to claim moral leadership on behalf of the global majority in the midst of a great power rivalry. A NITI Aayog AI strategy describes India as ' the AI Garage for 40% of the world ', suggesting that the AI use cases that India develops domestically will be exported to the global south.
3. For global development funders, like the Gates Foundation, who are pushing such initiatives elsewhere in the world under the label of AI for Development, the AI use case approach is the latest in a long line of digital interventions in development. It fits neatly within the philanthrocapitalist dogma that the solution to poverty is marketisation.
4. For the domestic software industry, the discourse of use cases legitimises the state-supported marketisation of a new digital population within India. It enables the extraction of citizens' data under the guise of development, though the financial value of these data and these customers is open to question. It offers the poor as test subjects in developing their products, while also opening up potential export markets in other developing countries.
5. For global tech giants, it offers a path to legitimise their activities in India. One of the most enthusiastic supporters of AI use cases is Microsoft CEO Satya Nadella, who recently (echoing Nilekani) remarked that India had become the 'AI use case capital of the world.' One of the most widely cited examples of AI in action for socio-economic development is the Jugalbandi chatbot, developed by Microsoft and IIT Madras, which provides vernacular language information about government services, and was released amidst a PR blitz in 2023. Adoption and usage statistics for the chatbot are unavailable. No further news has been released since 2023, and the project's website is no longer active.
The discourse of AI use cases is seductive because it poses an excellent question: Why shouldn't the poor benefit from the most advanced technology? Unfortunately, socio-economic development use cases as currently articulated won't succeed within the contemporary conjuncture. The hype is unlikely to benefit the poor or India's AI ambitions. It leaves dominant power structures undisturbed and doesn't challenge the monopolistic and extractive practices that undergird Big Tech-led AI. Instead, it is empowering a range of powerful actors.
What would it look like to centre poor and marginalised people while challenging Big Tech in an AI age? Most of all, it would require a shift away from treating people merely as end-users, data sources, and testing grounds of AI, but as its owners and producers. While genuine alternatives to the current set-up are largely speculative, initiatives imagining and working toward AI as a commons may provide inspiration for the kinds of changes that would be required for AI development to go hand-in-hand with socio-economic justice.
Mila T Samdub researches the aesthetics and political economy of digital infrastructure in India. He is a Visiting Fellow at the Information Society Project at Yale Law School, a CyberBRICS Fellow at the Center for Technology and Society, Fundacao Getulio Vargas, and an Open Future Fellow.
The article was first published in India in Transition , a publication of the Center for the Advanced Study of India, University of Pennsylvania.