AI Disinformation Spreads Quickly Before US Midterm Elections

As millions of Americans prepare to vote in the upcoming midterm elections, experts are warning about a dangerous new trend. Social media feeds are currently filled with a large amount of false information created by artificial intelligence (AI). This technology allows people to create fake images, videos, and stories very quickly and at a very low cost. According to The Guardian, this flood of misinformation is reaching levels that were impossible to achieve in past elections.

In the past, social media companies worked hard to protect voters from lies. However, the situation has changed significantly. After the Cambridge Analytica scandal, many companies built large teams to monitor content and ensure safety. But recently, these companies have reduced the size of these teams. The Guardian reports that this change happened partly because of pressure from political groups who worried that social media companies were limiting free speech. As a result, many platforms now rely on AI tools to catch lies instead of using human workers.

Frances Haugen, a former Facebook engineer who now works to make technology companies more responsible, explained why this is a problem. She noted that people who want to spread lies can now run campaigns that are very high quality but very cheap. Because they can change their posts slightly every time they share them, it is easy to avoid the rules that social media companies have in place to stop bad content. This means that many false posts are simply slipping through the cracks of the safety systems.

Political campaigns across the country, from Pennsylvania to Texas, are struggling to deal with these lies. When a false story appears online, campaigns often do not know who to contact to get it removed. The Guardian shared the story of James Talarico, a Democrat running for the US Senate in Texas. Earlier this month, he took a short break from his campaign because he had the flu. However, a false story soon appeared online claiming he was actually in a mental health facility in Ohio. This lie was supported by a fake, edited phone call with a hospital worker.

Even though Talarico’s team released video evidence showing he was at home in Texas, the lie continued to spread for several days. The Digital Democracy Institute of the Americas (DDIA) tracked how this story changed as it moved across the internet. They found that the lie even changed when it was translated into Spanish. In some Spanish-language media, the story changed from a claim about a hospital stay to a false report that the candidate had attempted suicide. This shows how misinformation can grow and change, making it very difficult for the truth to catch up.

Roberta Braga, the leader of the DDIA, noted that social media companies often do not enforce their rules as well in languages other than English. She explained that there are fewer protections in place for Spanish-speaking users. Many companies have moved toward a system where they rely on the community to report bad content, rather than hiring professionals to monitor it. This approach, which is sometimes called community-centered content moderation, often fails to stop lies before they cause damage.

When asked about these issues, major companies gave different responses. According to The Guardian, Meta—which owns Facebook and Instagram—and X did not answer requests for comment. YouTube, which is owned by Google, stated that all of its users must follow the same guidelines. However, experts remain worried that these rules are not enough to stop the current wave of AI-driven lies.

As election day gets closer, the challenge of fighting disinformation remains a major concern. Campaigns are finding that traditional ways of correcting the record are no longer enough. Because lies can change and mutate as they spread, a simple correction often fails to reach everyone who saw the original false story. For voters, this means it is becoming more important than ever to be careful about what they read and share on social media. The combination of easy-to-use AI tools and weaker safety rules on major platforms has created a difficult environment for everyone involved in the democratic process.