Global artificial intelligence models are failing to adapt sufficiently to local needs, and this presents a significant strategic opportunity for India, according to a leading academic. Professor Susan Athey, an expert in the Economics of Technology at Stanford Graduate School of Business, emphasised that for AI to be truly effective, it must solve problems "to the finish line," a process that requires deep collaboration with local implementing partners.
The Customisation Imperative for India
Speaking on the sidelines of a workshop on AI, jobs, and growth at the Delhi School of Economics in late December 2025, Professor Athey pointed out a historical trend. Global technology providers have consistently fallen short in customising their offerings to meet the specific domestic requirements of individual countries. This gap is particularly pronounced in India, where the need for adaptation spans language, local services, and data context. "The global models are often not providing that," Athey stated, identifying this as a clear opening for Indian enterprises and innovators.
She expressed increased optimism about India's capacity to become a major participant in the global AI ecosystem following her visit. The country's vast scale and diverse requirements mean that multinational corporations, often locked in a "race to the top" for general-purpose AI, lack the resources to serve every niche. This creates space for local providers who can build specialised application-layer solutions. "AI is not useful unless it's actually solving the problem to the finish line," Athey reiterated, highlighting that successful implementation demands integration with local systems, whether in healthcare, government, or defence.
Sovereignty, Regulation, and the Balance of Power
The professor also addressed critical concerns for developing economies about becoming net importers of AI technology, which could impact growth and national sovereignty. She warned of the risks of economic dependency on a single foreign entity. However, she noted a key leverage point for nations like India: AI derives its value from implementation within local systems—payment networks, healthcare data, and infrastructure—which are controlled domestically. This interdependence forces foreign vendors to cooperate.
Regarding India's draft framework on copyright and AI data, which proposes a blanket licensing regime, Athey advocated for a balanced approach. Any law must reconcile the needs of cash-strapped startups to access data with the imperative to compensate content creators like news organisations. "News involves research and development every day. It involves innovation every day," she said, equating journalistic reporting with AI innovation. Both forms of innovation, she argued, deserve protection and support within a fair regulatory structure.
The Role of Open Models and Future-Proofing
Professor Athey also commented on the rise of open-source AI models, such as China's DeepSeek, which challenge US dominance. She praised open models for enhancing competition and sovereignty, as they can be downloaded, customised with local language data, and retained indefinitely, preventing vendor lock-in or arbitrary price hikes. "Open models are really important because you can download them and keep them forever and so nobody can raise the price on you later," she explained.
However, she cautioned that while existing open models may suffice for many current applications, ensuring access to future cutting-edge advancements remains a complex challenge. For mission-critical integration into national infrastructure and essential services, India must develop a strategy for local control to mitigate risks of external disruption. The combination of a large domestic market, government-led demand, and the necessity for deep customisation positions India uniquely to build its own expertise and capabilities in the AI revolution.
