New Delhi, India – The Ministry of Electronics and Information Technology (MeitY) today launched AIKosha, a landmark initiative to create a comprehensive national repository of datasets and sandboxes for AI model development and evaluation. Unveiled at the India Mobile Congress 2025, AIKosha is designed to be a one-stop-shop for high-quality, curated, and contextually relevant Indian data, addressing one of the most significant hurdles for AI innovators in the country.
For years, Indian AI startups and researchers have struggled with the lack of large-scale, high-quality datasets that reflect India’s unique diversity. Most state-of-the-art AI models are trained on data from North America and Europe, leading to biases and lower performance when applied to Indian languages, cultures, and scenarios. AIKosha aims to solve this problem by providing a centralized platform for both public and anonymized private datasets that are critical for building effective and unbiased AI.
The platform will host a wide array of data types, including text corpora for dozens of Indian languages, labeled image datasets for applications in agriculture and retail, anonymized healthcare records for medical AI research, and speech data for voice recognition systems. The government will be seeding the repository with extensive public data and is creating a framework to encourage private companies to contribute anonymized data for the public good, with appropriate privacy and security safeguards.
Beyond being a simple data library, AIKosha will also feature secure sandboxed environments. These are controlled computational spaces where researchers and developers can access sensitive datasets to train and test their models without the data ever leaving the secure environment. This is particularly crucial for fields like healthcare and finance, where data privacy is paramount. By providing both the fuel (data) and the engine (secure sandboxes), MeitY’s AIKosha is set to become a cornerstone of the IndiaAI Mission, empowering a new generation of developers to build AI solutions that are truly made for India.
