Elon Musk proposes that major AI competitors conduct adversarial peer reviews of each other's models to improve AI safety, arguing that competitive oversight and product liability concerns would incentivize responsible development without requiring international regulation.