India is building a database for local companies to train AI models, MoS Electronics & IT

India is building a large database of anonymised non-personal data for Indian companies and startups that are using artificial intelligence (AI), said Rajeev Chandrasekhar, MoS for Electronics and Information Technology, at Global Fintech Fest (GFF), an industry event in Mumbai on Wednesday. 

Chandrasekhar said that it will be “one of the largest datasets programs where anonymised non-personal data will be available for the entire AI ecosystem.”

Though the database is primarily meant for the use of Indian companies, Chandrasekhar noted that an exception can be made for foreign companies that want to “come and create opportunities” in India. He also said that the government is weighing in favour of partnerships between Indian and foreign companies, particularly in deep tech. To encourage such partnerships, the government will also provide capital access including government contracts, he added.

Deep Tech is a classification used for startups or companies that are working on cutting-edge technologies for the future. 

Large databases play a key role in training machine learning (ML) models. Globally, there are various open-source datasets such as MySQL and Apache Cassandra, which are available to AI researchers and are being used by small and large companies across the world. 

Chandrasekhar further said that the government will soon introduce a Digital India Act, replacing the Information Technology (IT) Act 2000. Last month, the government withdrew the Personal Data Protection Bill, which received backlash from privacy advocates and tech companies for giving broad access to government agencies and for diluting it by mixing non-personal data with it. 

“We withdrew the Data Protection Bill 2021 because it had become too complex,” Chandrasekhar added. He reiterated that the government will soon bring a Digital Personal Data Protection Bill, which will be a “contemporary modern global standard framework of laws.”