Six months before the 2026 Tamil Nadu Assembly elections, a fabricated video using artificial intelligence set the political arena ablaze. The clip, which spread rapidly on social media, showed the state's revered first chief minister, C N Annadurai, apparently calling out to actor-politician Vijay with his iconic line, "thambi vaa, thalaimai yerka vaa" (come brother, take the lead). This marked one of the first widely reported instances of generative AI being deployed for political messaging in the state, instantly triggering a fierce ethical debate and criticism from rival parties.
The AI Arms Race Begins
The immediate and forceful political reaction to that initial deepfake established the template for AI's contentious entry into state politics. The Dravida Munnetra Kazhagam (DMK), which traces its ideological roots to Annadurai, retaliated within a day. Its IT wing released an AI-assisted video invoking Periyar and Annadurai, which was then amplified by party leaders accusing opponents of misappropriating the Dravidian legacy through technology. Political observer R Chandrasekaran notes that this exchange clarified that the objection was to the video's content, not the use of AI itself, signaling AI's transition from a fringe experiment to a mainstream campaign tool.
This episode proved to be a turning point. Parties quickly realized that AI could transform a single speech or image into countless social media posts much faster than traditional methods like WhatsApp forwards, posters, or manually edited videos. In the subsequent weeks, AI-assisted video production became routine for major players like the DMK, the AIADMK, and the BJP's Tamil Nadu unit.
While the BJP's state wing often adapted national AI content by swapping in local faces, the AIADMK leaned on animated videos, such as a viral clip about an 'Uruttu Kadai' (grinding shop) halwa, to accuse the DMK of broken promises. The DMK countered with videos like 'Pathu Tholvi Palaniswami' (Ten Mistakes Palaniswami), targeting AIADMK's Edappadi K. Palaniswami. "Once a party releases a provocative AI clip, we have to quickly respond with similar content to avoid falling behind online," explains Kovai Sathyan, AIADMK's IT wing state president.
From Campaigns to Character Assassination
The situation grew darker as AI use escalated beyond standard campaign messaging. A flood of deepfake videos emerged, showing leaders in absurd, comic scenarios. This soon devolved into more serious misuse, with obscene and defamatory deepfake images and audio clips targeting politicians across party lines. Faces were pasted onto explicit content, and fake audio was created to embarrass leaders.
"This soon gave way to more serious misuse, with obscene and defamatory deepfake images and videos targeting leaders from various parties," states R Gopinath, a political science academic. In response, parties began filing police complaints. While some arrests have been made, party insiders reveal that tracing creators is notoriously difficult. The content is often produced using easily accessible tools like Stable Diffusion for images, ElevenLabs for audio, and Reface or DeepFaceLab for videos, frequently by operators located outside India.
AI as the New Campaign Think Tank
Despite the murky side, parties are now leveraging AI for more than just mudslinging. Analysts predict this could be the first full election cycle in Tamil Nadu where AI plays a central, strategic role. The DMK, for instance, has announced an AI-powered portal to collect and categorize public suggestions for its 2026 manifesto.
"AI is increasingly becoming the 'think tank' behind campaigns," says Jai Pratap, director of Political Edge, a consultancy firm that has worked with major Tamil Nadu parties. He contrasts the old method of teams brainstorming for days with the new reality of prompting AI for instant campaign ideas. Jai Pratap explains that AI is now used to analyze demographic data, translate insights into targeted communication, replicate speeches, send personalized calls, and automate WhatsApp videos for occasions like birthdays and anniversaries.
He cites examples where AI designed a youth mobilization campaign in Chennai by first understanding the area's economic and social perspectives, and another where it built a candidate's strategy around the sitting MLA's perceived inaccessibility. His firm is also testing AI filters on Instagram and Facebook, allowing party workers to easily share branded content at scale.
The innovation extends to voter interaction. Divyendra Singh Jadoun, founder of The Indian Deepfakes, has developed AI voice-cloned calling systems that enable two-way communication. Voters receive a call that sounds like the candidate, which listens to their concerns and responds in real time, with data fed back to refine strategies. His firm even created an interactive AI avatar of former Andhra Pradesh Chief Minister Y S Rajasekhara Reddy.
However, strategists caution that key campaign decisions still rely on traditional surveys, field workers, and booth-level feedback. Off-the-shelf AI tools struggle to capture nuanced local factors like caste dynamics and hyper-local issues. Furthermore, parties are wary of incomplete or outdated voter data. "For now, parties see this technology as something that boosts their work, not replaces organisation on the ground," says a DMK war room staffer.
The Regulatory Lag and Future Trajectory
The breakneck speed of innovation has far outpaced regulation and detection capabilities. "Detection of fakes has become difficult," admits Divyendra Singh Jadoun. "What once took days and significant resources can now be created in minutes for under five dollars." He adds that the hyper-realistic outputs can sometimes confuse even experts.
Jadoun's company employs watermarks to denote AI-generated content and refuses to hand over tools directly to parties to prevent misuse. Jai Pratap confirms that some parties have requested services involving pornographic content or impersonation, which are firmly rejected. The focus for ethical consultants remains on image enhancement, not malicious manipulation.
Looking ahead, experts like D Raj, who has worked with political consultancies, point to global lessons. AI-generated robocalls mimicking US President Joe Biden's voice in the 2024 elections hint at a future where Tamil Nadu parties might shift from one-way IVRS calls to interactive voice systems in leaders' cloned voices. AI could also be used for sophisticated voter behavior analysis and volunteer tracking, similar to tools used by the Democratic National Committee in the US.
As the 2026 election approaches, the gap between AI's potential and the ground reality may narrow. But the journey from that first viral deepfake of Annadurai has already irrevocably changed the rules of political engagement in Tamil Nadu, setting the stage for a high-tech, and increasingly contentious, battle for votes.
