Swarms of artificial intelligence (AI) agents could soon invade social media platforms en masse to spread false narratives, harass users and undermine democracy, researchers warn. The next-generation AI swarms, on the other hand, are coordinated by large language models (LLMs) — the AI systems behind popular chatbots. Social media companies want real users on their platforms, not AI agents, so the researchers envisage that companies will respond to AI swarms with improved account authentication — forcing users to prove they are real people. Still, the researchers noted that strengthening authentication would make it more difficult and costly for those wishing to deploy AI swarms. "We need to be proactive instead of waiting for the first type of larger events being negatively influenced by AI swarms."