The government has superseded its previous AI advisory dated March 1 with a fresh notice and scrapped the much-critiqued "expicit approval" requirement by large platforms for rolling out "untested" AI models and LLMs.
In a fresh notice sent mainly to big tech platforms in India late on Friday, the Ministry of Electronics and Information Technology removed the requirement for obtaining approval and also submitting action-taken-cum-status reports.
However, it maintained its stance on labeling and the use of "consent popups" to inform users of the possible inherent "fallibility or unreliability" of the output generated by under-tested or unreliable AI models.
It further retained that the duty lies with the platform to ensure that there's no bias or discrimination in the output of the AI models or anything unlawful as well, it said.
In the advisory dated March 1, MeitY told all intermediaries or platforms that the use of undertesting or unreliable AI models, large language models or generative AI software "must be done so with explicit permission of the government of India," which witnessed pushback from various startups.
Senior IT Minister Ashwini Vaishnaw had to then clarify that the advisory was aimed only at large platforms such as Meta and Google and not Indian startups working in the realm of AI.
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.
