AI Disinformation Floods Feeds

Tens of millions of Americans face a surge of artificial intelligence-driven false narratives on social media just weeks before the November midterm elections. Experts describe the current volume of disinformation as a wave that is slipping past human reviewers. They argue that the speed and volume of these posts have outpaced the ability of platforms to correct errors in real time.

The shift in moderation strategy began after the Cambridge Analytica scandal in 2018. Following that event, major platforms strengthened trust and safety teams to protect voters. A subsequent political backlash grew after Elon Musk acquired X and Donald Trump won the 2024 presidential election. Fears of speech restrictions led to layoffs in safety divisions. Platforms now rely more heavily on automated tools than human staff to review content.

Frances Haugen, a former Facebook engineer, noted that the cost of running sophisticated information operations has dropped sharply. She stated that operatives can now achieve quality levels that were impossible in 2016 for a fraction of the previous cost. Because AI systems can adjust posts until they pass moderation filters, Haugen said some false content is guaranteed to reach voters.

Campaigns in key states are reacting to specific incidents of viral misinformation. In Texas, Democratic Senate candidate James Talarico was forced to address a false story claiming he was being treated at a mental health facility in Ohio. The claim originated from a deceptively edited phone call published by a right-wing provocateur. Talarico’s team released security camera footage showing him at home in Texas. The lie persisted for several days before Spanish-language media mutated the story into a false claim of a suicide attempt.

Roberta Braga, founder of the Digital Democracy Institute of the Americas, said platforms enforce rules less consistently in Spanish than in English. She added that increased reliance on AI moderation means even more errors are likely to slip through. Talarico’s Republican opponent, Ken Paxton, also used an AI-generated ad earlier this year depicting Talarico rewriting the Bible. Talarico told NBC News in September that such deepfakes do not add value to civic discourse.

In Pennsylvania, viral posts claimed an undocumented immigrant arrested in central Pennsylvania was a canvasser for Democrat Janelle Stelson’s congressional campaign. ICE deputy director Charles Wall publicly questioned the claim, calling it unbelievable that Stelson would employ an illegal alien. Stelson’s campaign provided records showing the man, Chimaobim Adedayo Igbojekwe, held no paid position. Republican congressman Scott Perry’s campaign continues to run ads containing the claim. Citizens for Responsibility and Ethics in Washington filed a complaint alleging a Hatch Act violation by the Department of Homeland Security.

Platform rules regarding election content remain narrow. Meta requires disclosure for AI content in election ads and bans posts that misstate voting dates or locations. It does not heavily moderate political commentary. YouTube stopped removing false claims about past elections in 2023. X replaced internal safety reviews with crowdsourced community notes. Nora Benavidez of Free Press said President Trump has used federal agencies to interfere in elections. Stelson’s campaign has not yet filed a direct complaint with the platforms but is considering legal avenues.

Haugen argued that moderating individual posts is insufficient when AI can tweak content to bypass filters. She said attention must turn to distribution systems and their inherent biases. These systems create feedback cycles that allow communities to either amplify or dampen harmful content.