AI-made ads cost about Rs 1,300 crore in US midterms, 81% of them Republican

Artificial intelligence (AI) is being used heavily in campaigning for the 2026 US midterm elections. Parties' digital work was earlier limited to making posters, writing and handling social media posts. Now AI tools are being used to create lifelike deepfake videos, fake conversations and misleading video ads to influence voters' minds.

A joint analysis by researchers at the Wesleyan Media Project and AdImpact, which tracks advertising spending, has found that more than 220 distinct political ads made with AI have appeared in this election so far. Running them on television channels, digital platforms and social media has cost about $137 million, roughly Rs 1,200 crore to Rs 1,300 crore. The scale of spending has raised concern worldwide over the growing intrusion of AI into politics.

According to the research data, 81% of these AI-made ads were commissioned by Republican candidates or political action committees (PACs) linked to them. The Democratic Party and its allied groups account for only about 12%. The Republican Party has thus made AI its main weapon to corner opponents and take its message to voters.

According to political analysis, AI is being used to damage opponents' images and to present a party's policies more attractively. AI can show any leader at events, or with leaders, with whom they have never had any connection. This has made it easier to confuse voters and sway their opinion.

Two recent examples of this campaign came from New York and Oklahoma. In New York, the campaign of Bruce Blakeman, the Republican candidate for governor, released a fully AI-made video. It showed New York Mayor Zohran Mamdani and Governor Kathy Hochul walking together and gardening in a park. The video advertised a fictional drug called the "Political Delusion Pill".

The video mocked the governor's and the mayor's policies on serious issues such as rising prices and crime. It looked so real that an ordinary voter could take it to be true. Most worryingly, the ad carried no disclaimer saying it was made with AI.

A similar case occurred in Oklahoma. A television ad backing a Republican candidate showed the candidate's opponent hugging Hillary Clinton and holding her hand. A message reading "Clinton lover, Trump hater" ran in the background, in an attempt to portray the candidate as anti-Trump. Running this fake AI video on television cost more than $2,000,000 (about Rs 1.6 crore), according to AdImpact.

Making video ads the old way meant choosing actors, shooting at different locations and editing. It cost lakhs of dollars and took weeks. With AI tools, any imaginary event or conversation can now be produced in a few hours at far lower cost. Parties are making different ads for voters in different areas, focusing on prices in some places, crime in others and refugees elsewhere. This is called micro-targeting.

AI is making campaigns faster and cheaper, but it is also posing a major challenge to democracy. Misinformation, fake videos and facts presented without context have made it hard for voters to tell truth from falsehood. Bringing transparency to political advertising and legally restricting AI-made content will be a major challenge for democracies in the US and across the world.