The Cyberspace Administration of China (CAC), in collaboration with four other departments, has jointly released the ‘Interim Measures for the Management of AI-Powered Human-Like Interactive Services.’ These measures are slated to officially come into force on July 15, 2026. Human-like interactive services involve leveraging artificial intelligence technology to emulate the personality traits of real individuals and offer ongoing emotional interaction services; services that do not meet this description are not governed by these measures. The ‘Measures’ underscore the equal significance of development and security, adopting a regulatory approach that is inclusive, cautious, and tailored to different categories. They outline promotional strategies and specify six types of activities that service providers are prohibited from engaging in. Concurrently, they mandate service providers to clearly define their primary safety responsibilities and implement systems such as safety evaluations and algorithm registrations to foster the creation of an AI sandbox safety service platform. A pertinent official from the CAC highlighted that the advancement and governance of human-like interactive services necessitate the collaborative involvement of multiple stakeholders across society.