AI-Powered Disinformation Swarms Are Coming for Democracy (www.wired.com)

A new paper published in Science highlights the evolving threat of AI-powered disinformation campaigns, suggesting that a single individual equipped with advanced AI tools could orchestrate vast "swarms" of social media accounts, capable of producing human-like content and adapting in real time. This shift marks a significant departure from previous disinformation tactics that relied on human-operated troll farms, as these AI agents would possess persistent identities and memory, allowing them to mimic human behaviors and engage with real users while coordinating their activities to amplify specific narratives. The implications for democracy are concerning, as these AI swarms could manipulate public opinion on a large scale, targeting specific communities with tailored messages. Experts warn that this new form of misinformation could disrupt future elections, particularly the 2028 presidential election in the United States, by creating an illusion of grassroots movements where none exist. To combat this looming threat, researchers advocate for the establishment of an "AI Influence Observatory" comprised of academics and non-governmental organizations, aiming to enhance awareness and response mechanisms, though they express skepticism about the willingness of social media companies to address these challenges given their profit motives.