India warns tech firms Need government approval before releasing ‘unreliable’ AI tools

The Indian government has cautioned tech firms creating new lab-level AI technologies that they must obtain approval before making them publicly available online.

India’s IT Deputy Minister, Rajeev Chandrasekhar, confirmed the development in a Tweet on Monday.

The development follows barely a week after a senior politician criticised Google’s Gemini AI tool on February 23 in response to accusations made by some that Indian Prime politician Narendra Modi was pursuing “fascist” policies.

In contrast to India’s prior laissez-faire attitude towards AI legislation, the new guideline calls for tech companies to “appropriately” indicate the “possible and inherent fallibility or unreliability” of the output that their AI models generate. 

The ministry decided against regulating the rise of AI less than a year ago, citing the industry’s criticality to India’s strategic interests.

Furthermore, because general elections are expected this summer, the warning asked platforms to ensure that their tools will not “threaten the integrity of the electoral process.”

Many business executives were surprised by the Indian authorities’ action. 

Numerous Indian businesses and venture capitalists express their fear and believe that the new guidance will make it more difficult for their country to compete in the global arena, where it is already falling behind.

Reacting to the development, Pratik Desai, founder of startup Kisan AI said: 

“I was such a fool thinking I will work bringing GenAI to Indian agriculture from SF,” 

“We were training multimodal low-cost pest and disease model, and so excited about it. This is terrible and demotivating after working 4 years full time bringing AI to this domain in India.”Subsequently, Chandrasekhar made it clear that the advice was only intended to “advise to those deploying lab-level or under-tested AI platforms onto public internet” about their legal responsibilities, the repercussions of their actions, and how to best safeguard both themselves and other users.