AI disinformation costs politicians days to correct before midterms

Tens of millions of Americans are set to vote in November, but experts say a surge of artificial intelligence-driven lies is cluttering social media feeds. Many of these posts are slipping through platform safety nets, creating confusion in the days leading up to the election. Social media companies have scaled back their human review teams, shifting the burden of detection to automated systems and community feedback tools.

Frances Haugen, a former Facebook data engineer and whistleblower, noted that the cost of running a high-quality information operation has dropped sharply. In 2016, such campaigns required significant resources. Today, operatives can tweak posts slightly to bypass moderation rules, allowing false narratives to persist. This shift has left campaigns scrambling to identify which platform or person is responsible for spreading a specific lie.

In Texas, Democratic Senate candidate James Talarico took time off from campaigning earlier this month to recover from the flu. While he was home, a false story spread across social media claiming he was being treated in a mental health facility in Ohio. A right-wing provocateur pushed the claim using a deceptively edited phone call with a hospital staffer. Talarico’s team released security camera footage showing him in Texas and clarified he had never visited Ohio.

The lie circulated for several days before it was corrected. The Digital Democracy Institute of the Americas tracked how the misinformation mutated as it moved through conservative Spanish-language media. The initial claim of hospitalization in Ohio shifted into a false report that Talarico had attempted suicide. Roberta Braga, the institute’s founder, explained that platforms enforce their rules more consistently in English than in other languages. With the move toward AI-driven moderation, fewer guardrails exist to catch these evolving falsehoods in Spanish.

Republican opponent Ken Paxton also used AI-generated footage in a campaign ad earlier this year. The video depicted Talarico, a Presbyterian seminarian, rewriting the Bible. Talarico told NBC News in September that Texans are tired of such lies. He argued that AI deepfake ads do not help voters understand the choices in front of them.

In Pennsylvania, Democratic congressional candidate Janelle Stelson’s campaign dealt with a viral claim that an undocumented immigrant arrested for indecent exposure was a paid canvasser for her race. An initial police report said the man, Chimaobim Adedayo Igbojekwe, claimed to be a campaign worker. Stelson’s team released records showing he had no paid position. However, Charles Wall, the deputy director of Immigration and Customs Enforcement, amplified the claim in an official statement. He called it unbelievable that Stelson would employ an illegal alien for the campaign.

The Stelson campaign sent cease-and-desist letters to media outlets and activists filed a complaint with Citizens for Responsibility and Ethics in Washington. The complaint accused the Department of Homeland Security of violating the Hatch Act, which limits the political activities of government employees. The campaign of Stelson’s opponent, Republican Scott Perry, continues to run ads featuring the claim. A spokesperson for Stelson said the campaign is considering every legal avenue to fight the false claims.

Platforms maintain specific rules for election misinformation. Meta requires disclosure of synthetic media in ads and bans posts that misrepresent voting dates or locations. YouTube stopped removing content that falsely claimed past elections had widespread fraud in 2023, citing a chilling effect on speech. It now relies on information panels to amplify authoritative sources. X replaced internal safety reviews with a crowdsourced community notes program after Elon Musk’s acquisition. Experts argue that without strict enforcement, these policies often fail to stop the spread of AI-generated lies.