Abstract: Artificial intelligence legislation has long been trapped in the tripartite balance paradox of efficiency,security and innovation.Confined to the linear logic of value trade-offs,the existing regulatory approaches struggle to break through this dilemma.The essence of law lies in a construction mechanism of social trust.By clarifying rights and obligations,defining boundaries of conduct,and guaranteeing remedy for rights,law delivers stable expectations,which inherently aligns with the governance demands of the artificial intelligence era.Trustworthiness is not a fourth dimension independent of the three core values;instead,it serves as the logical hub that integrates the three to achieve dynamic coordination.Establishing trustworthiness as the core objective of artificial intelligence legislation can not only resolve the inherent contradictions of the traditional balance model,but also address trust crises such as algorithmic black boxes and data abuse.Based on this core proposition,a dedicated artificial intelligence law should be formulated in the future.With trust building as its legislative purpose,it shall establish a set of principles including trust construction,hierarchical and classified regulation,scenario adaptation and multi-party collaboration.Through refined institutional design covering trustworthy algorithms,credible data,accountable entities and sound governance,it will provide steady value guidance for the rule of law in artificial intelligence and foster positive interaction between technological application and social trust.