AI deepfakes are no longer only a threat discussed in terms of fake scandals or manipulated clips. In India’s general election, they have also become a practical campaign instrument, used by politicians and service providers to scale voter outreach across a vast and multilingual electorate.
The source article follows Divyendra Singh Jadoun, a self-taught deepfaker whose firm, Polymath Synthetic Media Solutions, builds AI versions of politicians for calls, videos and campaign media. His work shows how quickly synthetic media can move from novelty to routine political infrastructure.
A campaign message built from a candidate’s voice
On an April afternoon in Ajmer, in the Indian state of Rajasthan, local politician Shakti Singh Rathore sat for a greenscreen recording session. He wore a white shirt and a ceremonial saffron scarf with the lotus flower, the logo of the BJP, the country’s ruling party.
Jadoun was not trying to make a normal campaign video. He needed enough audio and video material to create an AI deepfake of Rathore that could make 300,000 potential voters around Ajmer feel as if they had received a personalized message from him.
The technical requirements shaped the performance. Rathore had to face the camera and limit his movement, because too much motion would interfere with the algorithm. The recording itself mattered less as a finished speech than as raw material for a synthetic version of the politician.
Polymath Synthetic Media Solutions uses this kind of data to translate political speech into different languages for voter outreach. In Rathore’s case, the generated speech appeared in Hindi, Tamil, Sanskrit, and Marathi.
Why deepfakes fit India’s election machine
India’s scale is central to the story. Close to a billion Indians are eligible to vote in the general election, and the country has 22 official languages and thousands of regional dialects. For campaigns trying to reach voters far from major political centers, language and distance are not minor obstacles.
The source describes political parties across the ideological spectrum using AI for this outreach. The same technology often associated with misinformation, disinformation and social harm is being used to deliver customized political contact at scale.
That use creates a different kind of risk. A voter receiving a voice clone call or AI-generated video may hear a familiar political figure speak directly to them by name. The message can discuss issues that seem personally relevant, while the person on the other end may not realize they are interacting with a machine.
The business is already large. In India, sanctioned deepfakes have become a $60 million business opportunity. More than 50 million AI-generated voice clone calls were made in the two months leading up to the start of the elections in April, with millions more expected during voting, according to one of the country’s largest business messaging operators cited by WIRED.
The economics of synthetic campaigning
Jadoun’s company is one of many deepfake service providers working for India’s political class. During this election season, he had delivered five AI campaigns and had been paid a total of $55,000.
His prices show why the technology is attractive to campaigns. The source reports that he charges 125,000 rupees [$1,500] to make a digital avatar and 60,000 rupees [$720] for an audio clone, significantly less than big political consultants.
The services go beyond simple voice messages. Jadoun has made deepfakes for Prem Singh Tamang, the chief minister of the Himalayan state of Sikkim. He also resurrected Y. S. Rajasekhara Reddy, an iconic politician who died in a helicopter crash in 2009, to endorse his son Y. S. Jagan Mohan Reddy, currently chief minister of the state of Andhra Pradesh.
He has also created AI-generated propaganda songs for several politicians, including Tamang, a local candidate for parliament, and the chief minister of the western state of Maharashtra. One Hindi song about a local politician in Ajmer included the lines, “He is our pride” and “He’s always been impartial.”
Another campaign song for local politician Ram Chandra Choudhary in Ajmer included translated lyrics such as: “For Ajmer, he brought a new gift / His name is Ram Chandra / He helps everyone / He was the president of Ajmer Dairy / He has always been impartial.”
Personalization replaces the doorstep
Rathore is not up for election this year, but he is part of a much larger political apparatus. He is one of more than 18 million BJP volunteers working to help the government of Prime Minister Narendra Modi maintain power.
In the past, that work meant traveling across Rajasthan, a desert state roughly the size of Italy, to meet voters one by one. Volunteers would remind people about BJP social programs, including pensions, free tanks for cooking gas, and cash payments for pregnant women.
With synthetic media, that outreach can be partly automated. Rathore planned to spend 15 minutes answering prompts on camera while Jadoun collected what he needed. After that, the data could be turned into videos and calls sent directly to voters’ phones.
The intended experience is intimate. Instead of a handshake or a visit, a voter may hear or see Rathore address them by name, discuss relevant issues, and ask them to vote for the BJP. If voters ask questions, the AI is meant to respond in a clear and calm voice.
Rathore himself admits he does not know much about AI. But he understands the political goal. As he put it, “Such calls can help with swing voters.”
A political operator becomes an AI operator
Jadoun’s path into this industry came through politics. In 2015, he stood for election in Ajmer as district president of the National Students Union of India (NSUI), the youth wing of the Indian National Congress, the chief opposition to Modi’s BJP.
His campaign style was built around visibility. He hired four shotgun-wielding gunmen to travel with him in open-top jeeps, hoping to project power. “When you have gunmen with you, others start feeling envious,” he said. “Wherever I would go, these gunmen would come with me.”
He won that election, and two years later became the state general secretary of the NSUI for Rajasthan. But after five years in politics, he felt constrained by the demands around him. “I used to get calls at 2 am to handle situations,” he says. “I realized that my time and my life weren’t mine any longer.”
The Covid lockdowns gave him distance from politics and time to experiment. Stuck at home, he began a series of “30-day challenges” to learn new skills, including playing the flute, mastering graphic design, trading on the stock market, and becoming ambidextrous.
That background matters because the current deepfake boom is not just a technology story. It is also a campaign story, built by people who understand political performance, local networks and voter psychology. In India’s election, AI deepfakes are not merely distorting politics from the outside. They are being invited in by the candidates and operators who see them as a new way to campaign.