Hybrid world models enable synthetic reality

World models that combine large language models (for scripting and logic) with photoreal simulation engines and real‑time audio create editable, persistent synthetic scenes that users can control interactively. When offered by major platform owners as subscriber features, these systems can produce lifelike media, dialogues, and environments on demand. — This raises immediate public concerns about deepfakes, platform concentration of cultural production, safety/verification, monetization of simulated human interaction, and the need for provenance rules.