A swarm of ten AI agents can now outmaneuver a hundred human trolls and nobody sees it coming
New research from Cornell Tech and Ben-Gurion University demonstrates that a swarm of just ten LLM-powered AI agents can replicate the influence output of a hundred human trolls while evading virtually all standard bot detection systems. The study simulated a successful attack on an online voting platform and sparked over 50,000 social media interactions within 24 hours, raising urgent questions for platform operators, AI providers, and election security officials worldwide.