Researchers Warn AI Swarms Threaten Democracy

A peer-reviewed study published in Science on Feb. 1, 2026 warns that LLM-coordinated "AI swarms" could mimic human behavior on social media, spread misinformation, and harass dissenting users. The authors model how swarms could manufacture consensus and propose defenses—including stronger account authentication, live-traffic anomaly detection, and an "AI Influence Observatory"—to help platforms and regulators mitigate threats to democratic discourse.