### Policy Reversal: Anthropic Goes From Blacklisted to Landing a CIA Contract The field of artificial intelligence is witnessing a dramatic turning point. Anthropic, an AI startup that was previously placed on the U.S. government's "Entity List" due to data security controversies, has recently been officially awarded a $230 million contract by the Central Intelligence Agency (CIA) to develop an intelligent analysis system. This policy reversal marks a fundamental shift in the U.S. government's regulatory logic regarding AI technology—moving from "risk-first" to "capability-first." Sources indicate that the contract requires Anthropic to build a specialized large language model (LLM) system for the CIA capable of processing multi-source intelligence and performing real-time semantic analysis. In a statement, CEO Dario Amodei emphasized, "This is not just a commercial victory, but an official endorsement of the feasibility of our AI safety framework." Industry analysts point out that this move could reshape the landscape of the AI industry: on one hand, compliance capability has become a core threshold for enterprises seeking government contracts; on the other hand, the AI arms race is deepening from the consumer market into the realm of national security. It is worth noting that Anthropic's "Constitutional AI" safety architecture may become the standard paradigm for future government-level AI projects.