India’s finance ministry has banned employees from using artificial intelligence tools including ChatGPT and DeepSeek for official work, citing risks to government data security.
The directive comes just as OpenAI’s chief executive officer Sam Altman visits India for talks with government officials.
The advisory, dated 29 January, explicitly warns about the confidentiality risks posed by AI applications. “It has been determined that AI tools and AI apps (such as ChatGPT, DeepSeek etc.) in the office computers and devices pose risks for confidentiality of (government) data and documents,” the finance ministry stated in its internal directive.
Three finance ministry officials confirmed the authenticity of the note, which was circulated internally this week. The move aligns India with countries like Australia and Italy, which have implemented similar restrictions due to data security concerns.
The timing is particularly notable as it coincides with Altman’s first visit to India since 2023. During his visit, Altman met with information technology minister Ashwini Vaishnaw to discuss potential collaborations in AI development.
Vaishnaw shared details of the meeting on social media platform X, describing it as a “super cool discussion” about India’s strategy to develop its AI infrastructure. The minister highlighted India’s ambitions to create a cost-effective AI ecosystem, drawing parallels with the country’s successful space programme.
“Our country sent a mission to the moon at a friction of the cost that many other countries did right, why can’t we do a model that will be a fraction of the cost that many others do?” Vaishnaw noted during his discussion with Altman.
The restrictions come at a challenging time for OpenAI in India, which represents its second-largest market by user numbers. The company is currently embroiled in copyright infringement disputes with several of India’s leading media organisations and has argued in court filings that Indian courts lack jurisdiction as it does not maintain servers in the country.
OpenAI and DeepSeek are yet to comment on the advisory. It remains unclear whether similar restrictions will be implemented across other Indian government departments.
