India Orders 3-Hour Takedowns

India’s Election Commission has told social media platforms to take down unlawful election posts, including fake or AI-made content, within three hours during the 2026 state polls. (medianama.com) The directive applies to the 2026 assembly elections in Assam, Kerala, Tamil Nadu, Puducherry and West Bengal. Medianama reported the order on April 22, and other reports said the Commission had already acted on more than 11,000 violative posts and links. (medianama.com, msn.com) The three-hour deadline is not entirely new. In May 2024, the Press Information Bureau published an Election Commission direction telling political parties to remove fake content within three hours of learning about it and warning against AI deepfakes in campaigns. (pib.gov.in) A separate rulemaking track in New Delhi is pushing platforms beyond takedowns and into product design. The Indian Express reported on April 22 that the Ministry of Electronics and Information Technology wants AI-generated posts to carry “continuous and clearly visible” labels for the full duration of the content. (indianexpress.com) Hindustan Times reported the same proposal as an “always-on” label requirement for synthetic media on social platforms. That would replace earlier compliance that officials said was too weak or inconsistent when labels appeared briefly, faintly or not at all. (hindustantimes.com, indianexpress.com) The government’s draft AI-content rules have been moving in stages since October 2025. The Ministry of Electronics and Information Technology said its draft amendments were meant to address “synthetically generated information,” a category it tied to deepfakes, misinformation, impersonation and election manipulation. (meity.gov.in) Those synthetic-content amendments were formally added to India’s IT Rules on February 10, 2026, according to the ministry’s updated rules page. A separate March 30, 2026 draft then proposed tighter intermediary-compliance obligations, and the ministry extended consultation on that draft on April 10. (meity.gov.in, meity.gov.in) For platforms, the combined effect is operational: faster complaint handling during elections, plus labels that stay on-screen instead of appearing once and disappearing. That forces moderation, reporting and publishing systems to track whether a post is unlawful, whether it is synthetic, and whether the disclosure remains visible from start to finish. (medianama.com, indianexpress.com, hindustantimes.com) India is not banning AI campaign media outright. It is building a system that demands two things at once during an election: remove unlawful fakes fast, and keep lawful synthetic content visibly marked the whole time people watch it. (pib.gov.in, indianexpress.com)