AI political falsehoods spread online ahead of US midterms
Experts warn that AI-generated political falsehoods are spreading widely on social media before November’s US midterm elections, while platforms have reduced investment in voter protection and content moderation. The combination makes it easier to produce convincing material cheaply and harder for campaigns to find effective ways to correct it.
In Texas, a false claim that Democratic Senate candidate James Talarico was receiving mental health treatment in Ohio spread for days, then mutated in Spanish-language media into a claim that he had attempted suicide. Researchers say corrections may not cover each altered version, and moderation is less consistent in Spanish than in English. Talarico’s Republican opponent, Ken Paxton, has also used AI-generated footage of him in a campaign advert.
In short
- Experts report a surge in AI-driven election disinformation.
- A false story about James Talarico changed as it spread.
- Researchers cite weaker moderation of Spanish-language content.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Platforms and campaigns have a responsibility to limit deceptive election material because realistic falsehoods can spread quickly, mutate across languages and outpace corrections, potentially misleading voters. The reported inconsistency in Spanish-language moderation strengthens the case for investing in safeguards and making corrections easier to find.
The case against
Heavy moderation can also suppress legitimate political speech, satire or material that is mistaken for false, especially when claims change quickly and context is disputed. A more cautious approach would prioritise clear labelling, transparent rules and timely corrections, while avoiding broad removals that could unfairly shape campaign debate.
Explained simply
US midterm elections take place halfway through a president’s four-year term. Voters choose members of Congress, as well as some state and local officials, and the results can affect which party controls Congress and what laws can pass.
Social media platforms are major channels for political news and campaign messages. AI tools can make realistic-looking images, audio and video, while false claims can be copied and altered quickly; content moderation means platforms’ efforts to review or limit such material.
In Texas, James Talarico and Ken Paxton are competing for a US Senate seat. False claims about a candidate can reach voters in different forms and languages, making them harder to track and correct.
Americas Elections Politics Technology World
Read the full article at the source →
Originally published by The Guardian as “AI-driven disinformation running rampant on social media before US midterms, experts warn”.