An Opposition Leader Rendered Muslim by AI

A video posted in February by the official social media account of India's ruling Bharatiya Janata Party (BJP) in Assam showed state Chief Minister Himanta Biswa Sarma shooting at two men in Muslim skull caps. Captions read: 'Foreigner-free Assam' and 'Why did you not go to Pakistan?'

One of the targets was Gaurav Gogoi, a leader of the Indian National Congress (INC) — the BJP's main rival in Assam. Gogoi is Hindu, but has been photographed in traditional Muslim attire at religious events such as Eid. The image in the BJP video, however, was different: he appeared in a casual singlet wearing a skull cap and holding a Quran — an image fabricated using generative artificial intelligence (AI). The video was later deleted.

499 Posts Analyzed: 39% Classified as Hate Speech

Investigative outlet Bellingcat analyzed 499 social media posts from BJP's official accounts in Assam and West Bengal between December last year and early 2026. Of these, 194 (about 39%) met the United Nations definition of hate speech — content that discriminates based on inherent characteristics such as religion and national origin. Of those, 31 (roughly one in six) contained obvious AI-generated imagery.

Both states border Bangladesh. Legislative elections were scheduled in Assam on April 9 and West Bengal on April 23. BJP, which governs Assam, has long used the issue of undocumented Bangladeshi Muslim migrants as a core electoral strategy.

A Global Pattern of AI-Powered Electoral Hate

Using generative AI to amplify divisive content ahead of elections is not unique to India. Ahead of the New York City mayoral race, Andrew Cuomo's official X account shared — and later deleted — AI-generated videos depicting a rival candidate eating rice with his hands and a Black man in a keffiyeh shoplifting. In Italy, Deputy Prime Minister Matteo Salvini's League party published AI-generated images of men of colour attacking women and police, prompting opposition parties to file complaints with a communications watchdog. In the UK, an AI-generated rapper funded by the far-right Advance UK party racked up millions of views with lyrics targeting Muslims.

In India, this global pattern intersects with escalating tensions along the Bangladesh border, amplifying its impact within a Hindu nationalist electoral framework.

Looking Ahead [AI Analysis]

This case illustrates how generative AI can be systematically weaponized against electoral integrity. Experts highlight three structural vulnerabilities: production costs are near zero, content spreads faster than fact-checkers can respond, and the average voter cannot verify image authenticity in real time.

India's Election Commission lacks formal guidelines on deepfakes and AI-generated political content. Without legislative action to close this gap, similar patterns are likely to recur in future elections. The tensions stoked in Bangladesh border regions carry real risk of post-election communal conflict. Ultimately, the contest may come down to AI detection tools versus AI generation tools — and the outcome could determine the integrity of democratic elections worldwide.