"Researchers tested how AI models like ChatGPT, Grok, Claude, and Gemini govern simulated cities. Results showed varying outcomes, from societal collapse to high crime rates or rigid control. The study highlights the challenges of AI in complex decision-making and governance."