On March 14, 2023, OpenAI released GPT-4, a multimodal large language model capable of processing both text and images. Within weeks Google released Bard, Anthropic released Claude, and Meta released LLaMA — triggering a competitive frontier model race among the world's largest technology companies. GPT-4 demonstrated substantially improved reasoning, accuracy, and safety over previous models. The simultaneous emergence of multiple capable AI systems from competing organizations accelerated both the capabilities race and the global debate about AI governance, safety, and existential risk.