Indian government requires approval for new AI tools before launch

The Indian government has taken an important step by advising technology companies working on developing new artificial intelligence (AI) tools that these tools must first be approved by the government before they can be made publicly available.

This advisory, issued on March 1 by the Ministry of Information Technology, specifically targets AI tools that are considered “unreliable” or still in the testing phase. Such tools should be clearly marked to indicate that they may be inaccurate in answering questions.

The ministry emphasizes that the availability of these tools on the Indian internet requires explicit permission from the Indian government. This measure appears to be partly motivated by the need to guarantee the integrity of the electoral process, in view of the general elections that will take place later this summer.

This new directive follows an incident in which a top Indian minister criticized Google and its AI tool Gemini for providing “inaccurate” or biased answers, including a description by some of Indian Prime Minister Narendra Modi as a “fascist” . Google has apologized for this, saying the tool “may not always be reliable,” especially on contemporary social topics.

Unreliability does not exempt you from the law

Deputy IT Minister Rajeev Chandrasekhar emphasized that “security and trust are legal obligations of a platform” and an excuse of unreliability does not mean exemption from the law. Earlier, the Indian government had announced that it would introduce new regulations to combat the spread of AI-generated deepfakes ahead of the elections.

These guidelines have received criticism from the technology community, which argues that India is at the forefront of the technology world and it would be a mistake for the country to place itself outside this leading position. Chandrasekhar responded to this criticism by saying that India is not giving up, but rather embracing its ambition in AI and a safe and reliable internet for its users.

These developments come amid a collaboration between Microsoft and Indian AI startup Sarvam to bring an Indic-Voice Large Language Model to Azure’s AI infrastructure, which will increase accessibility to more users in the Indian subcontinent.