India vs AI risks: Breather for big digital platforms on AI models deployment, but here’s here’s the catch

The government has announced the elimination of the approval mandate for major social media subsidiaries and internet giants prior to the launch or deployment of their AI models within India.

This decision, heralded by the Ministry of Electronics and IT, comes as a revision to its AI advisory issued on March 1. The essence of this update is that prominent internet entities are now liberated from the necessity of seeking governmental authorisation before introducing any AI model in India.

However, this newfound freedom comes with a caveat. Big digital platforms are encouraged to conspicuously label any AI models deemed "under-tested and unreliable" to apprise users of their potential limitations or flaws.

According to the fresh directive from the Ministry, any under-tested or unreliable AI models, foundational or otherwise, must be presented to Indian users with clear labels highlighting their inherent fallibility or unreliability.

Moreover, all intermediaries and platforms must ensure that the utilisation of AI models, be it for generative AI or any other software or algorithms, does not facilitate the hosting, display, modification, or dissemination of unlawful content, as stipulated in Rule 3(1)(b) of the IT Rules or any other provisions of the IT Act.

The advisory underscores the potential penal repercussions, including prosecution under the IT Act and other relevant statutes, for intermediaries, platforms, or users found in violation of these regulations.

The digital platforms are instructed to promptly adhere to the new AI guidelines, effective immediately.

Furthermore, it has been clarified by the Centre that this relaxation does not extend to startups, as they are not covered under the permission to launch new AI models.

The government has articulated that the objective of this advisory is to prevent untested AI platforms from infiltrating the Indian internet landscape, underscoring its commitment to ensuring responsible and accountable deployment of AI technologies.