India is changing its AI stance and requiring government approval for model launches

India has waded into the worldwide AI debate by issuing an advisory requiring tech firms to hunt government approval before launching latest models.

India’s Ministry of Electronics and IT issued the advisory to corporations on Friday. The opinion, which isn’t publicly available but a duplicate of which is reviewed by TechCrunch, also calls on tech corporations to make sure that their services or products “don’t permit bias or discrimination or jeopardize the integrity of the electoral process.”

Although the ministry admits the suggestion isn’t legally binding, India’s deputy IT minister Rajeev Chandrasekhar says the notice “signals that that is the long run of regulation.”

He added: “We are making this today as a suggestion and asking you to stick to it.”

In its statement, the ministry refers back to the powers conferred on it by the IT Act, 2000 and the IT Rules, 2021. It seeks compliance with “immediate effect” and requires tech corporations to submit an “motion with status report” to the ministry inside 15 days.

The latest notice, which also calls on tech firms to “appropriately” label the “possible and inherent fallibility or unreliability” of the outcomes generated by their AI models, marks a turnaround from India’s previous, hands-off approach to AI regulation. Less than a 12 months ago, the ministry had declined to manage AI growth, as an alternative identifying the sector as critical to India’s strategic interests.

India’s move has surprised many industry executives. Many Indian startups and VCs say they’ve been unsettled by the brand new advisory and imagine such regulation will impact the country’s ability to compete in the worldwide race, where it’s already lagging behind.

“I used to be silly enough to think I’d work to bring GenAI from San Francisco to Indian agriculture.” wrote Pratik Desai, founding father of the startup Kisan AI. “We trained a multimodal, low-cost pest and disease model and were very enthusiastic about it. This is terrible and demotivating after working full-time for 4 years to introduce AI on this field in India.”

Many Silicon Valley executives also criticized India’s political changes. Aravind Srinivas, co-founder and managing director of Perplexity AI, one in every of the most popular AI startups, said the brand new Delhi-based consultancy is a “bad move by India.”

Martin Casado, partner on the enterprise firm Andreessen Horowitz, said, “Good rattling sir. What a travesty.”

The reference comes after Chandrasekhar expressed disappointment over a certain response from Google’s Gemini last month. Last month, a user asked Gemini, formerly often called Bard, whether India’s Prime Minister Narendra Modi was a fascist.

In response, Gemini – citing experts it didn’t discover – said Modi had been accused of implementing policies that some had described as fascist. Chandrasekhar responded to the exchange by warning Google that such responses were “direct violations” of IT Rules 2021 in addition to “several provisions of the Penal Code.”

Failure to comply with the provisions of the IT Act and IT Rules would have “potential criminal consequences for the intermediaries or platforms or their users in the event that they are identified,” the advisory further states.