The AI Safety Bargain: Why Washington Is Betting on Voluntary Red Teams

A U.S. approach built on soft law and adversarial testing puts evaluation at the center of AI governance, while the EU keeps moving toward binding compliance.