
Three strategies policymakers are weighing to avert an AI catastrophe
Researchers warn rogue, superintelligent AI could threaten humanity, with Anthropic highlighting a meaningful probability within a decade; lawmakers and researchers discuss three routes: slow development with international cooperation, stronger government guardrails, or tighter in-house restrictions on AI capabilities. Each option faces challenges—from competitive risks with China to enforcement gaps and risk of unintended actions as models grow more capable.












