The ongoing debate within the industry about the rapid advancement of AI has captured global interest, causing notable volatility in the stock prices of AI-related entities in Japan and South Korea. Several internationally acclaimed AI firms have come to an agreement, urging for a worldwide cap on the research and development of state-of-the-art AI technologies to lessen the threat of AI spiraling out of control. In reaction, 360 articulated that merely imposing a speed limit would not eradicate the risks at their core. Instead, it emphasized that AI security necessitates the integration of independent third-party offensive and defensive capabilities to address the shortcomings of self-assessment. Furthermore, 360 put forward an AI security approach termed as 'leveraging models to oversee models,' championing the creation of a more robust security governance framework grounded in a third-party independent offensive and defensive verification system. This proposal offers fresh perspectives for the global governance of AI security.