Abstract: Development and risks are two sides of the evolution of things. Out of the strong requirements of national security, international competition, domestic economic growth and social development, the policy system determines the direction of the development of AI and provides supports through pressure transmission, planning guidance and other means. Capital actively responds to government decisions, integrates production factors under the guidance of market competition and self-value-added mechanisms, and promotes artificial intelligence technology research and development to commercial use. With policy encouragement and capital support, artificial intelligence technology continues to iterate and mature along the path of“self-evolution”. In this power chain that connects the former and the later, the power that comes back to bite the achievements of technological development is also imperceptibly growing. Through the“technology-capital-policy”reverse cycle path, artificial intelligence technology risks continue to overflow and spread. Under the combined effect of increased technological autonomy, intensified capital expansion, and excessive policy intervention, endogenous technical security vulnerabilities may evolve into technical misuse, abuse, and misuse in the field of production services, leading to serious deviations from the established goals of the policy. These changes will bring greater uncertainty and challenges to the healthy development of national security, economy, and society. In the process of vigorously advancing innovation-driven development, when further releasing the potential of artificial intelligence, to prevent and resolve major scientific and technological risks has become a governance strategy and principle that need to be adhered to for a long time.

Key words: artificial intelligence, intelligent society, digital risks, driven by innovation