The pervasive risk of bias in generative artificial intelligence (AI) systems necessitates robust measures to protect public rights and enhance regulatory effectiveness. Addressing bias across the lifecycle of AI products—from data collection and training to modeling and application—requires legal and technical strategies tailored to each layer of potential bias. Current regulatory frameworks, such as China's “Interim Measures for the Administration of Generative AI Service,” lack specific guidelines to mitigate bias in AI decision-making. Global regulatory frameworks are still developing, underscoring the need for a comprehensive governance structure that defines the scope of regulation, implements layered measures to address bias, and allocates liability among platform developers.