Following the completion of the most extensive election in the world on June 5, 2024, where more than 640 million ballots were tallied, onlookers had the opportunity to evaluate the utilization of artificial intelligence technologies by the different political parties and factions—highlighting lessons applicable to other regions.
The electoral campaigns heavily relied on AI, incorporating sophisticated imitations of candidates, personalities, and deceased politicians through deepfake technology. It is estimated that several million Indian voters encountered deepfakes.
Despite concerns about widespread misinformation, for the larger part, the electoral campaigns, candidates, and activists used AI positively during the election period. AI was employed not only for traditional political maneuvers, such as mudslinging, but mainly to establish stronger connections with the electorate.
Deepfakes minus the deceit
Political parties in India invested approximately US$50 million in approved AI-generated content for targeted communication with their voter bases in this recent election cycle, a strategy that resulted in considerable success.
Indian political strategists have long acknowledged the impact of individuality and sentiments on their supporters, prompting the adoption of AI to enhance their messaging. Emerging AI enterprises like The Indian Deepfaker, which initially catered to the entertainment sector, swiftly responded to the surging demand for AI-generated campaign material.
In January, Muthuvel Karunanidhi, the former long-time chief minister of Tamil Nadu in the southern region, made an appearance via video at his party’s youth division conference. Despite his passing in 2018, his party authorized the deployment of a deepfake showcasing Karunanidhi with his trademark yellow scarf, white shirt, dark shades, and characteristic pose.
In February, the official X account of the All-India Anna Dravidian Progressive Federation party shared an audio recording of Jayaram Jayalalithaa, the iconic political figure in Tamil Nadu affectionately known as “Amma” or “Mother,” who had passed away in 2016.
Simultaneously, constituents received calls from their local representatives to discuss community concerns, though the speaker on the other end of the line was actually an AI impersonation. Individuals like Shakti Singh Rathore from the Bhartiya Janta Party (BJP) frequently engaged with AI startups to send personalized video messages to specific voters via WhatsApp, informing them about government benefits they had received and soliciting their electoral support.
Enhanced Multilingual Support
Deepfakes were not the singular application of AI in the Indian elections. Well before the election kick-off, Indian Prime Minister Narendra Modi spoke to a packed audience celebrating the connections between Tamil Nadu in the south and Varanasi in Uttar Pradesh in the north. Introducing his audience to wear headphones, Modi proudly introduced his “new AI technology” as he delivered his speech in Hindi, being real-time translated into Tamil.
In a nation boasting 22 official languages and nearly 780 unofficial recorded dialects, the BJP integrated AI tools to make Modi’s persona accessible to voters in regions where Hindi comprehension was limited. Since 2022, Modi and the BJP have harnessed the AI-driven platform Bhashini, integrated within the NaMo mobile application, to translate Modi’s speeches with voiceovers into Telugu, Tamil, Malayalam, Kannada, Odia, Bengali, Marathi, and Punjabi.
As part of their presentations, some AI firms circulated distinctive versions of Modi’s renowned monthly radio program “Mann Ki Baat,” meaning “From the Heart,” which they tailored to regional languages using voice replication.
Utilization in Contradictory Contexts
Indian political factions heightened online trolling, employing AI to amplify their ongoing meme battles. Early into the election period, the Indian National Congress unveiled a brief clip to its six-million Instagram followers, featuring the theme song from a new Hindi music album titled “Chor” (thief). The clip fused Modi’s digital likeness onto the lead vocalist and mimicked his voice with revised lyrics criticizing his close relationships with Indian business magnates.
In retaliation, the BJP released its own clip on its seven-million-follower Instagram account, showcasing a montage of Modi campaigning on the streets intertwined with footage of his supporters against the backdrop of unique musical composition. The music was an old patriotic Hindi song sung by the renowned artist Mahendra Kapoor, who passed away in 2008, but was synthetically revived using AI voice emulation.
Modi himself acknowledged an AI-generated video of him dancing in a quote-tweet, a common meme format that juxtaposes footage of rapper Lil Yachty on stage, remarking “such creative content during the peak of the election season is truly enjoyable.”
In certain instances, the inflammatory language in Modi’s campaign that endangered the Muslim community and incited violence was disseminated through generative AI mechanisms, yet the responsibility lies with the hateful discourse itself rather than the AI instruments utilized to propagate it.
The Indian Case Study
India stands as an early adapter, and its venture into AI offers a representation of what could unfold in forthcoming elections worldwide. Although AI’s capacity to generate non-consensual deepfakes complicates the differentiation between truth and falsehood, its consensual applications are likely to enhance the accessibility of democracy.
The integration of AI during the Indian election, from the realms of entertainment to political meme warfare, emotional appeals, resurrected figures, and personalized voter calls, has established a path for AI’s involvement in participatory democracy.
The unexpected outcome of the election, where the BJP failed to secure the anticipated parliamentary majority, heralds India’s return to a highly competitive political landscape, emphasizing the potential for AI to play a constructive role in deliberative democracy and representative governance.
Implications for Global Democracies
In a democratic setting, it is imperative for political entities and candidates to establish targeted interactions with their voter base. The Indian elections exemplify a distinctive endeavor to leverage AI for personalized outreach across linguistically and culturally diverse constituencies, fostering greater accessibility especially among rural and low-income populations.
The future of participatory democracy in conjunction with AI could transform constituent communication from mere personalization to an actual dialogue, enabling voters to voice their demands and experiences directly to their representatives instantaneously and on a large scale.
India can be a beacon of extending its recent proficiency in AI-assisted party-to-people communications beyond politics. The administration is already utilizing these platforms to deliver government services to citizens in their native tongues.
If harnessed conscientiously and ethically, this technology could present an opportunity for a new era in representative governance, with a particular focus on ensuring the voices and realities of rural populations reach the national legislature.
This composition was jointly crafted with Vandinika Shukla and was initially published in The Conversation.
About Author
