Amitabh Kant Says India Must Build AI on Domestic Data to Prevent Tech Dependence

New Delhi, Feb 17: Amitabh Kant on Tuesday said India and other countries in the Global South should build their own artificial intelligence (AI) models using local data to ensure that the technology drives inclusive growth and does not exacerbate inequality.
Speaking at a panel discussion during the India AI Impact Summit 2026 here, the former NITI Aayog CEO stressed that AI systems must be multilingual, accessible, affordable and accountable.
He cautioned that the rapid pace of AI development and the scale of investments flowing into the sector could lead to major societal disruption and create a highly unequal society if not designed inclusively.
“The challenge is whether we can ensure that AI reaches the population which is below the poverty line, whether it can be used to transform the lives of citizens in the Global South and whether it can be used to improve learning, health outcomes, and nutritional standards,” Kant said.
He underlined that AI must benefit people below the poverty line and be leveraged to improve education, healthcare and nutrition outcomes across developing nations.
Highlighting the importance of linguistic diversity, Kant said, “If AI is not multilingual, it will cut out a large section of the population.”
Referring to large language models (LLMs), he noted that India contributes 33 per cent more data than the United States for training such systems, and that models are increasingly improving using data generated in the Global South.
Kant warned that global technology firms could build business models around such data and later sell AI-driven products at high costs to the very countries that contributed to their development. To prevent this, he said India and other developing nations must create their own AI models based on domestic datasets to ensure equitable access and benefits.
Drawing parallels with India’s experience in building digital public infrastructure, Kant said the country had leapfrogged decades of development through open and interoperable systems enabled by open APIs and global interoperability.
This approach, he said, spurred innovation in fast payments, stock market transactions, insurance and last-mile credit delivery, while allowing the private sector to build and compete on top of a robust public digital backbone.
He suggested adopting a similar model for AI by creating a digital public identity layer on which private players can innovate.
As AI adoption accelerates globally, Kant emphasised that the technology must be deployed to improve learning outcomes, enhance healthcare delivery and raise nutritional standards, ensuring that growth in the Global South remains inclusive rather than exclusionary.