AI executives are deliberately creating myths about existential risks and AGI to consolidate power and extract resources, according to investigative journalist Karen Hao's research of 250+ OpenAI insiders. Her book reveals how these companies operate as extractive empires—seizing intellectual property, exploiting labor, and sacrificing vulnerable communities—while gaslighting the public about technological capabilities and progress. OpenAI systematically defines AGI differently depending on audience (existential threat to Congress, digital assistant to consumers, $100B revenue generator to investors) to manipulate policy and capital allocation without coherent vision. Data annotation has become a major job category absorbing displaced workers in inhumane conditions where they train models to automate the exact roles they lost, creating perpetual instability and breaking career advancement pathways.