Nitin Upadhyay
We are framing the India AI story within a broader developmental paradigm, not a competitive or geopolitical one. In the larger international discourse, AI is framed in terms of frontier model performance, risk of regulation, and strategic competition between major economies. Nations with concentrated capital, existing research infrastructure, and large-scale computational capacity largely shape such discourses and others like them. Their natural policy preoccupations are, therefore, safety, liability, intellectual property, and control. While these are valid considerations, they do not adequately capture the realities of large, diverse, and developing societies. Therefore, we must understand the emerging AI trajectory in India within the context of a distinct social and economic reality.
In a society characterized by linguistic diversity, regional imbalance, and large public service needs, AI cannot remain a remote technological aspiration. Its relevance is in its ability to be integrated with education, healthcare, agriculture, and governance. The conversation about AI policy in India is increasingly framing AI as a tool for a bigger institutional change agenda, rather than as a race for technological leadership. The question, therefore, is no longer merely about the sophistication of the models but about their ability to overcome the constraints of everyday life and provide greater access to opportunities.
There are three interconnected elements that form the foundation of any AI ecosystem: compute, data, and inference. Each of these is a technical element, but it is driven by economic and policy considerations. The first of these is compute. This is one of the most visible global bottlenecks. Specialized GPUs and cloud computing are expensive and often controlled by a few corporations. Such constraints can make innovation accessible only to well-funded parties, which can also decrease the diversity of experiments. In a nod to this structural issue, India has started investing in shared compute resources through initiatives like the IndiaAI Mission. The idea is to increase access for universities, startups, and government agencies, thereby decreasing reliance on third-party sources.
The second pillar—data—poses even more fundamental questions related to representation and context. AI systems use the strength of the data they receive for training. As a matter of fact, many of the most successful models around the world are English-language or Western-originated data sets, which may not be representative of the cultural and linguistic diversity of other parts of the world. India, with its many languages and different social and economic situations, is a good example of this. If AI systems are to be effective in such environments, they need to be trained on data sets that are representative of the local environment.
This aspect has led to an increasing focus on the development of curated and multilingual data ecosystems. Projects such as BHASHINI under the National Language Translation Mission is geared to improve speech recognition, translation, and text processing capabilities for Indian languages. The IndiaAI Kosh platform aims to provide the broader community with structured data sets and fundamental models.
Alongside growth, there is an increasing concern with issues of privacy, data protection, and ethics. Instead of viewing data as merely a resource for extraction, concerns about stewardship and sovereign control are growing. The aim is not simply to accumulate data but to ensure that it remains available for public interest innovation while also respecting individual rights.
Inference—the third element—applies compute and data to applications. Sector-wise applications in the Indian context reflect this aspect. Learning platforms that conform to the National Education Policy (NEP) 2020 are incorporating adaptive learning solutions that help teachers and provide customized learning experiences. The healthcare sector is testing AI-based diagnostic platforms to detect diseases like tuberculosis early, particularly in remote regions. Agricultural advisory services are employing predictive analytics to help with crop planning and management. Although the development of overall impact assessments continues, there appears to be a gradual transition from testing to incorporating these technologies into public systems.
However, governance is a key factor in this transition. The regulatory framework in India has generally sought to strike a balance between encouraging innovation and developing a framework related to fairness, transparency, and accountability. Instead of imposing a strict framework from the beginning, the government has taken a balanced approach, using a combination of legislation and sector-wise guidelines. Often this approach allows for flexibility in adapting to the evolution of these technologies, it, however, also demands constant monitoring to ensure that they are not misused.
The other dimension of the Indian approach is its focus on digital public infrastructure. Examples of how interoperable platforms can scale up and reduce transaction costs include the Aadhaar platform and the Unified Payments Interface. The AI ecosystem seems to be following suit, considering artificial intelligence an extension of publicly accessible digital infrastructure. Policymakers are attempting to ensure that there is not an undue concentration of capability in a few private companies.
Thus, the Indian approach embodies a desire to balance ambition with social responsibility. On the one hand, there is an acknowledgement that AI will impact economic competitiveness and state capacity. On the other hand, there is a realization that unregulated technological concentration may worsen inequality. Through simultaneous investments in compute access, multilingual datasets, sectoral deployment, and adaptive governance, India is trying to develop an approach that is more attuned to its development needs.
The future course of this trajectory is still unclear. We will require heavy investments, coordination, and capacity building. Issues of digital divides, cybersecurity, and technical talent will remain. However, the developing framework provides a new approach to the global AI discourse. It indicates that leadership may not only be measured by the size of models or the magnitude of capital inflows but also by the degree to which AI systems are attuned to linguistic diversity and public service needs.
In this respect, the developing Indian model may have applicability for other countries that face similar structural challenges. By integrating artificial intelligence with development strategy and public infrastructure, it provides a new approach that emphasizes inclusion, relevance, and shared capability. Whether this focus will ultimately transform global norms will depend on its implementation. However, it is part of a larger rethink of how AI can be developed and governed in societies characterized by diversity, size, and complex development challenges.
Let this vision be anchored in “SAMARTH”—an Indian ethical commitment that power must remain accountable to society and that innovation fulfils its highest purpose when it uplifts every life it touches. Rooted in this commitment, SAMARTH calls for service to the common good through accountable institutions and resilient infrastructures, mission-driven innovation aligned with national priorities, secure and inclusive digital foundations, principled governance, and robust techno-legal stewardship, all while keeping human well-being at the centre of every technological advance. In this framework, Artificial Intelligence is not an end in itself but a transformative public instrument—deepening human agency, expanding opportunity, and advancing shared progress responsibly.DD
