AI Models, Including GenAI Models, Must Get Govt Nod Before Deployment: MeitY

MeitY has purportedly requested AI platforms to include metadata if the content generated by the AI could potentially propagate misinformation or produce deepfakes.

Union minister Rajeev Chandrasekhar on Monday gave insights into the recent advisory.

The Ministry of Electronics and Information Technology (MeitY) has issued an advisory stating that any artificial intelligence (AI) models, including generative AI models, undergoing testing or deemed unreliable in any aspect, must obtain permission from the government before deployment in the country, the media has reported. This advisory follows recent incidents where users discovered inaccuracies and misleading information in responses from Google's Gemini AI chatbot regarding Prime Minister Narendra Modi.

As reported by The Economic Times, the advisory was issued on March 1, instructing companies to adhere to its guidelines moving forward. The advisory urged firms that have already deployed AI platforms in the country to guarantee that their computer resources do not exhibit bias, discrimination, or jeopardise the integrity of the electoral process.

Furthermore, MeitY has purportedly requested AI platforms to include metadata if the content generated by the AI could potentially propagate misinformation or produce deepfakes.

Union minister Rajeev Chandrasekhar on Monday gave insights into the recent advisory issued by the Ministry of Electronics and Information Technology (MeitY) for Artificial Intelligence (AI) platforms. He clarified that the advisory exclusively pertains to major platforms and does not encompass startups in the country, as per Business Standard.

Taking to social media platform X, formerly Twitter, the minister underscored: "The recent MeitY advisory should be interpreted accurately. It specifically targets prominent platforms, and the requirement for permission from the ministry applies solely to large platforms, exempting startups."

Additionally, companies were asked to include clear disclaimers if their platform is prone to unreliable behaviour and can produce inaccurate information. Moreover, platforms will be required to caution users against utilising AI to generate deepfakes or any content that could influence elections in any manner, according to the report.

Although the advisory does not hold legal enforceability at present, it suggests the direction of AI regulation in India, moving forward.