AI4Bharat, supported by IIT Madras, and Josh Talks AI announced on Tuesday the launch of Voice of India—the country's first independent multimodal platform for evaluating artificial intelligence (AI) systems, designed to test AI performance in real Indian conditions.
This platform represents a new element in India's AI ecosystem. Despite significant national investments in language technologies, computing infrastructure, and AI models, an independent infrastructure for conducting such evaluations remained a critical component.
The goal of Voice of India is to become a reliable benchmark for AI evaluation in the country. This will allow governments, enterprises, researchers, and AI developers to independently verify how AI systems function considering India's linguistic diversity, including various languages, accents, and dialects, as well as in different deployment environments.
Listen to this article
Professor Mitesh Khapra from AI4Bharat at IIT Madras noted that India has a unique opportunity not only to create AI for the world but also to help define how the world evaluates it. He stated: "Evaluation determines what will be created. As AI systems integrate deeper into society, rigorous evaluation becomes as important as model development itself. Through Voice of India, we hope to provide the scientific community, industry, and policymakers with evidence-based evaluation frameworks that reflect India's linguistic diversity."
Instead of relying solely on performance metrics or benchmarks developed for Western contexts, the platform offers scientifically rigorous, India-specific evaluations that reflect the use of AI across the country.
The Voice of India AI Evaluation Platform was created through a collaboration between AI4Bharat at IIT Madras and Josh Talks AI. Its purpose is to serve the broader AI ecosystem in government, enterprise, academia, and industry. It is designed as an open, neutral, and scientifically rigorous platform, allowing AI developers to compare their systems while helping buyers make more informed decisions about adoption.
As part of the launch, AI4Bharat and Josh Talks AI presented a set of benchmarks within the Voice of India framework, including a finance-oriented benchmark to assess AI performance in key banking workflows.
Supriya Paul, co-founder of Josh Talks AI, emphasized: "India has a once-in-a-generation chance to contribute something unique to global AI—perhaps not the largest models, but an evaluation infrastructure that makes AI trustworthy for the entire world. The Voice of India AI Evaluation Platform is a step towards building that future."
The significance of this launch is that businesses and government agencies implementing AI in India can now, for the first time, compare models using independent, deployment-specific evaluations instead of relying solely on vendor benchmarks.
The platform introduces a three-stage evaluation system, which includes assessing individual models, end-to-end deployment evaluation in real working conditions, and periodic post-deployment audits. This allows organizations to assess not only how AI works in controlled environments but also how it functions for real users in real scenarios.
