Guardian: AI disinformation floods 2026 US midterm races

The lie: Texas Democratic Senate candidate James Talarico was hospitalized at an Ohio mental health facility. The reality: he had the flu. In Spanish-language versions circulating online, the fabrication mutated into a claim he had attempted suicide. The Guardian traces that case and others like it through platforms whose trust and safety operations have been scaled back since Elon Musk acquired Twitter and Donald Trump won the 2024 election.

Frances Haugen, the former Facebook data engineer and whistleblower, puts the economics plainly. "You can run an information operation now with quality that was impossible in 2016 for super, super cheap," she says. Moderation is easy to slip past, she adds, "because you can just keep tweaking the post a little bit each time you create them, and some of them are going to get through."

The examples the paper documents are not subtle. Talarico's Republican Senate opponent Ken Paxton has run campaign ads featuring AI-generated footage of Talarico, a Presbyterian seminarian, rewriting the Bible. In Pennsylvania, Democratic congressional candidate Janelle Stelson was falsely accused of employing an undocumented immigrant, Chimaobim Adedayo Igbojekwe, as a campaign worker; ICE deputy director Charles Wall amplified the claim.

Roberta Braga, who founded the Digital Democracy Institute of the Americas, says the gaps are widest for non-English content. "There's a lot less guardrails in place from the companies to deal with Spanish language," she tells the paper. Nora Benavidez of Free Press frames the state's role flatly: "Trump has leveraged agencies across the federal government, whether it's the US Postal Service or DHS."

Two of the researchers we follow had circulated the piece by the end of the day it ran.