India changes its stance on AI, now mandating government approval for model launches

India has entered the global AI discussion by issuing an advisory that mandates “significant” tech companies to obtain government approval before releasing new models.

An advisory was issued to firms by India’s Ministry of Electronics and IT on Friday. The advisory, which has been reviewed, requests that tech companies ensure that their services or products do not allow bias, discrimination, or compromise the integrity of the electoral process.

Even though the advisory is not legally binding, India’s IT Deputy Minister Rajeev Chandrasekhar suggests that it indicates the direction of future regulation. He states, “We are issuing an advisory today, requesting your compliance.”

Chandrasekhar mentioned in a tweet on Monday that the advisory targets AI platforms that have not been tested and are being deployed on the internet in India. He clarified that the advisory does not affect startups.

According to the ministry, it is empowered by the IT Act, 2000, and the IT Rules, 2021, as stated in its advisory. The directive requires immediate compliance from tech companies, requesting them to submit an “Action Taken-cum-Status Report” to the ministry within 15 days.

The recent advisory calls on tech companies to label the potential limitations of the AI models they produce, signaling a shift from India’s previous approach to AI regulation. Just under a year ago, the ministry decided against regulating AI growth, recognizing the sector’s importance to India’s strategic interests.

India’s recent decision has caught many industry executives off guard. Several Indian startups and VCs express concern over the new advisory, fearing it will impede the nation’s competitiveness in the global arena, where it is already falling behind.

Pratik Desai, founder of startup Kisan AI, expressed regret for his ambitious plans to introduce GenAI to Indian agriculture from SF. We were exploring a multimodal, low-cost pest and disease model and were extremely enthusiastic about it. This situation is disheartening and discouraging after dedicating four years full-time to introducing AI to this field in India.

Several Silicon Valley executives also expressed disapproval of India’s policy change. Aravind Srinivas, co-founder and chief executive of Perplexity AI, expressed his disapproval of the recent advisory from New Delhi, calling it a negative decision by India.

Martin Casado, a partner at venture firm Andreessen Horowitz, exclaimed, “Good heavens!” This is a complete disaster.

Following Chandrasekhar’s recent disappointment with a response from Google’s Gemini last month, an advisory has been issued. Last month, a user inquired of Gemini, formerly known as Bard, about whether India’s PM Narendra Modi was a fascist.

Gemini responded by mentioning unidentified experts who accused Modi of implementing policies described as fascist. Chandrasekhar issued a stern warning to Google, stating that their responses were in direct violation of the IT Rules, 2021, and several provisions of the Criminal Code.

Failure to adhere to the regulations outlined in the IT Act and IT Rules could lead to penalties for intermediaries, platforms, or their users upon detection, according to the advisory.