In Lok Sabha election campaigning, political parties put AI to test

Technology is being used to translate speeches instantly, create digital anchors and in one case make videos of a long-dead leader

As campaigning for the Lok Sabha elections heats up, political parties are using artificial intelligence (AI) to spread their message: Left, Right and centre. The technology helps in translating speeches instantly, creating digital anchors and making videos of a long-dead leader.

Prime Minister Narendra Modi used Bhashini, an AI-driven tool, to translate his speech in Hindi to a Tamil-speaking audience at an event in Varanasi in December. “I am using AI technology for the first time and in the future I will use it,” he said.
 

Months later, Modi is seen asking people at his election rallies to visit the dedicated regional language handles on X (formerly Twitter) and YouTube to listen to his AI-translated speeches. The Bharatiya Janata Party (BJP) has created YouTube channels that play Modi’s translated speeches in Bangla, Tamil, Odia, Malayalam and other languages.
 

Campaign mode
 

Like the BJP, its rivals the Congress and Dravida Munnetra Kazhagam (DMK) are using AI to create lifelike audiovisual media for campaigning. DMK, which governs Tamil Nadu, is using AI-generated videos that show former chief minister and party icon M Karunanidhi seeking voters’ support. Karunanidhi died in 2018 at age 94.
 

In West Bengal, the Communist Party of India (Marxist) is using an AI-generated anchor called Samata to create videos for the party’s social media outreach in election campaigning.

A Congress spokesperson said in March it plans translating party leader Rahul Gandhi’s speeches into different languages using GenAI and machine learning.
 

“Now, every party wants interesting, interactive content with AI. It is helping the parties to bring the aura of old leaders back in the form of their voices and popular slogans, and is being used both in a positive manner but also in satirical form,” said Niranjan R, founder of Political Analytics India (PA-I), a consultancy.
 

For the attention it is getting, AI’s use in elections is nascent and experimental. “Bigger political parties are still a little sceptical about using artificial intelligence on a large scale. It started as a great idea but issues of lack of proper expertise, overselling, and mediocre quality content have been a spoil-sport,” said Diggaj Mogra, a political analyst.
 

Hype about the technology has led to an overabundance in AI-generated content against real demand, said Mogra. Every political party is getting offers of similar AI-generated content like cloned videos, images and video files.
 

Studios and creative firms creating political content see elections as a trial run for the use of AI and are waiting for feedback on the technology later.
 

Divyendra Singh Jadoun, the founder of The Indian Deepfaker, which describes itself as a synthetic media company, said AI is actually being used on a small scale in campaigning. The reason could be that political parties “don’t want the content out already because it might give an edge to the opponent, in terms of either using it against them or producing similar content,” he said.
 

Senthil Nayagam, founder of Muonium AI, a Chennai-based visual effects studio, said, “Elections in Tamil Nadu are a week away (on April 19), we are done with making the content. It is up to them (political parties) on how they use that content now. We are yet to get a voter reaction on most of the content,” said Muonium, which created the lifelike digital version of Karunanidhi.
 

Other than regulatory concerns and voters’ view, AI-generated content has technical issues. “AI is not very good at creating Indian faces. We are expanding our datasets, and we are trying to improve in the next few months,” said Nayagam. “No political party would want to offer content that ends up backfiring on them or used against them by the opposition.”
 

The downside
 

The downside of AI is a concern for governments across the world, especially during elections. 
 

The risk is the technology being used to spread misinformation. AI-fuelled misinformation is the biggest threat to elections in India, said the ‘Global Risk Report, 2024’ by the World Economic Forum.
 

Big technology firms have taken steps against the spread of manipulated content after deep fake videos and images created a controversy and prompted the government to issue advisories.
 

Meta, which owns WhatsApp, Instagram and Facebook, will from May start labelling videos with “Made with AI” tag to help spot manipulated content. In a blog post in April, the company said that it has expanded its third-party fact-checking network in India and now has 12 partners for 16 Indian languages and English. 
 

Meta started an India-specific election command centre last month to fight misinformation and improve transparency across its platforms. On Instagram, the company has restricted political content ahead of elections in different countries.
 

Search giant Google has announced restricting Gemini, its AI platform, from answering queries on Indian elections. YouTube, the video-sharing platform owned by Google, recently introduced a tool which mandates creators to disclose to users if a piece of content was generated using synthetic media or with the help of generative AI. 
 

YouTube  has removed more than 7.3 million election ads by advertisers who did not complete verification in 2023.

X, on April 4, announced starting its Community Notes programme in India. It allows users to participate in fact-checking tweets.

 

AI’s poll test

AI-generated content in election campaigning is still limited, experimental

It could be that parties are cautious, are waiting to use the tech fully later

AI-generated content has issues, isn’t ‘good at creating Indian faces’

AI’s potential for misuse has prompted Big Tech to announce new policies