AI-generated images and video have become a prominent feature of the midterm election cycle, adding to the challenge for Instagram, YouTube and X as they try to monitor political content. The trend has been dubbed “slopaganda,” a term coined by AI researchers to describe Generative AI combined with propaganda and misinformation.
A Wesleyan Media Project analysis found that more than $187 million had been spent on nearly 300 political ads generated or enhanced with AI. The researchers said the total is likely an undercount because many ads lack disclosures. More than 90% of the spending came from Republican candidates, a Republican party committee or a Republican-leaning group, and most of the ads did not include disclosures, the report found.
There is no comprehensive federal law governing the use or disclosure of AI in political advertising. Around 30 states have laws with varying requirements, but candidates do not always follow them. AI-generated ads and posts from politicians are appearing alongside authentic content on platforms operated by Meta, Google and X.
Examples include an undisclosed fake video of New York Governor Kathy Hochul posted by her Republican opponent, Bruce Blakeman, according to CNBC. In Texas, Republican Senate candidate Ken Paxton created an AI-generated video portraying Democratic opponent James Talarico making controversial statements. The video includes a disclosure that appears for three seconds and has been shared by Paxton’s campaign as organic content on YouTube and X. Neither campaign immediately responded to requests for comment.
YouTube said it applies prominent labels to realistic AI content. The company also said political candidates can request removal of organic videos in which their face appears altered or AI-generated, while advertisers must disclose synthetic or digitally altered people or events in election ads. Google requires political advertisers to complete its election-ad verification process and include “paid for by” disclosures.
X says advertisers are responsible for complying with applicable laws. Meta pointed to its political-ad policies, including restrictions during the week before an election, its “paid for by” labels and its political-ad library. In June, Meta also introduced an “about this ad” tool for ads it detects as having been made with third-party AI tools.
Research by the Brennan Center for Justice found that several AI models could be manipulated into generating misleading material, even when they avoided producing content related to familiar conspiracy theories. In an October retest, ChatGPT had begun redirecting prompts for harmful election images toward comic-style illustrations or refusing them.
Bipartisan senators have backed the Protect Elections from Deceptive AI Act, which would prohibit knowingly distributing materially deceptive AI-generated video or audio about federal candidates. Other proposed legislation, including the AI Ads Act, would extend existing restrictions on campaign misrepresentation to generative AI content. None of the bills has advanced out of committee.
Sources: CNBC Markets.