Blab AI
Government Contracts, Not Laws, Shape AI Governance in India, Study Finds
In India, the real architect of artificial‑intelligence governance is not a statute but the clauses scrawled in procurement contracts. A recent study from the Human Sciences Research Center at IIIT‑Hyderabad shows that the country’s approach to AI is forged in tenders, not in the drafting of a dedicated AI Act.
The paper, titled Emerging Institutional Pathways for AI Governance in India, earned the Best Paper Award at the 6th India Public Policy Network Conference. It was written by dual‑degree student Sujal Deoda, researcher Siddhi Wadekar, and Professor Aakansha Natani.
Unlike the European Union, which adopted its AI Act in 2024, India still lacks a standalone legislative framework for the technology. The IIIT‑Hyderabad team set out to map where AI governance actually takes root in a regulatory landscape that is still evolving.
To answer that question, the researchers combed through publicly available tender documents, procurement notices, technical specifications, and contractual requirements issued by central and state government agencies for AI systems. Their analysis spanned a range of ministries and departments that regularly solicit AI‑driven solutions.
The study found that procurement documents frequently embed explicit requirements on standards, accountability, compliance, and performance. These clauses shape how AI systems are designed, tested, and deployed. Because the government rarely builds AI in‑house, it outsources the bulk of these services through tenders, making the operational choices made during procurement the main drivers of AI governance.
"The government needs AI‑driven technologies, but may not always have the capacity or expertise to build them internally. As a result, many of these services are outsourced through tenders," said Sujal Deoda, co‑author of the study.
Professor Natani added that a full picture of AI governance requires looking beyond the technology itself. "Researchers need to examine not only technical specifications but also what is written, and sometimes not written, in policy and procurement documents," she said.
The authors argue that as AI adoption spreads across public services, governance will increasingly emerge from everyday processes—commissioning, procuring, and deploying technology—rather than from statutes alone. They emphasize the need for interdisciplinary work that blends computer science, public policy, and social sciences to understand AI’s broader societal impact.
Their findings underscore procurement’s power as a lever for shaping AI governance. By embedding standards, accountability, and performance metrics into contracts, government agencies can influence the ethical and technical characteristics of the solutions they acquire.
While the study does not prescribe specific policy changes, it suggests that a more explicit regulatory framework could complement existing procurement practices. Such a framework might clarify vendor and agency responsibilities and offer clearer guidance on compliance with national and international standards.
The research arrives at a pivotal moment. India’s AI market is projected to reach $8 billion by 2025, and the government has launched initiatives such as the IndiaAI Mission to nurture the sector. Yet, without a dedicated AI Act, procurement contracts will likely remain central to how AI is governed in public institutions.
In short, the IIIT‑Hyderabad study demonstrates that the practicalities of buying AI solutions—through tenders and contracts—are currently the main drivers of AI governance in India. The authors call for a coordinated approach that brings together procurement policy, technical standards, and ethical considerations to ensure that AI deployed in government services aligns with the country’s social and economic objectives.
The study’s authors plan to continue exploring how procurement practices can be leveraged to promote responsible AI deployment and to engage with policymakers about potential regulatory frameworks that align with the realities of public procurement.
The paper, titled Emerging Institutional Pathways for AI Governance in India, earned the Best Paper Award at the 6th India Public Policy Network Conference. It was written by dual‑degree student Sujal Deoda, researcher Siddhi Wadekar, and Professor Aakansha Natani.
Unlike the European Union, which adopted its AI Act in 2024, India still lacks a standalone legislative framework for the technology. The IIIT‑Hyderabad team set out to map where AI governance actually takes root in a regulatory landscape that is still evolving.
To answer that question, the researchers combed through publicly available tender documents, procurement notices, technical specifications, and contractual requirements issued by central and state government agencies for AI systems. Their analysis spanned a range of ministries and departments that regularly solicit AI‑driven solutions.
The study found that procurement documents frequently embed explicit requirements on standards, accountability, compliance, and performance. These clauses shape how AI systems are designed, tested, and deployed. Because the government rarely builds AI in‑house, it outsources the bulk of these services through tenders, making the operational choices made during procurement the main drivers of AI governance.
"The government needs AI‑driven technologies, but may not always have the capacity or expertise to build them internally. As a result, many of these services are outsourced through tenders," said Sujal Deoda, co‑author of the study.
Professor Natani added that a full picture of AI governance requires looking beyond the technology itself. "Researchers need to examine not only technical specifications but also what is written, and sometimes not written, in policy and procurement documents," she said.
The authors argue that as AI adoption spreads across public services, governance will increasingly emerge from everyday processes—commissioning, procuring, and deploying technology—rather than from statutes alone. They emphasize the need for interdisciplinary work that blends computer science, public policy, and social sciences to understand AI’s broader societal impact.
Their findings underscore procurement’s power as a lever for shaping AI governance. By embedding standards, accountability, and performance metrics into contracts, government agencies can influence the ethical and technical characteristics of the solutions they acquire.
While the study does not prescribe specific policy changes, it suggests that a more explicit regulatory framework could complement existing procurement practices. Such a framework might clarify vendor and agency responsibilities and offer clearer guidance on compliance with national and international standards.
The research arrives at a pivotal moment. India’s AI market is projected to reach $8 billion by 2025, and the government has launched initiatives such as the IndiaAI Mission to nurture the sector. Yet, without a dedicated AI Act, procurement contracts will likely remain central to how AI is governed in public institutions.
In short, the IIIT‑Hyderabad study demonstrates that the practicalities of buying AI solutions—through tenders and contracts—are currently the main drivers of AI governance in India. The authors call for a coordinated approach that brings together procurement policy, technical standards, and ethical considerations to ensure that AI deployed in government services aligns with the country’s social and economic objectives.
The study’s authors plan to continue exploring how procurement practices can be leveraged to promote responsible AI deployment and to engage with policymakers about potential regulatory frameworks that align with the realities of public procurement.
