The U.S. is scrambling to strengthen guardrails around increasingly powerful artificial intelligence models before China can catch up. It may already be running out of time. New AI models, such as Anthropic’s Claude Mythos and OpenAI’s GPT 5.5-Cyber, have advanced faster than legislation regulating the technology can keep pace. They have both shown a remarkable…