[Guangdong has registered 195 large models, ranking among the top in the country] On October 17th, the results conference of the Guangdong-Hong Kong-Macao Greater Bay Area Generative Artificial Intelligence Security Development Joint Laboratory was held in Guangzhou on October 16th. A number of results were released at the meeting. The "Guangdong Province Registered Large Model Observation Report (2026)" systematically sorts out the development status, structural characteristics, and innovation highlights of the current 195 registered large models, providing reference guidance for improving industry governance policies and optimizing the industrial development environment; the "Joint Laboratory Technology System Construction Plan (2026-2028)" proposes the accelerated evolution of generative artificial intelligence from large models to agents and large-scale applications. The new trend will focus on promoting capacity building such as security assessment, situational awareness, agent and application security assessment, trustworthy assessment, cross-border governance collaboration between Guangdong, Hong Kong and Macao, and AI security think tanks; the "APEC Artificial Intelligence Legal Policy Blue Book (2026)" conducts a comparative study on the artificial intelligence legal policies of all 21 APEC member economies, and makes suggestions for Guangdong to seize opportunities and deepen regional coordination. The meeting also introduced Guangzhou’s achievements in promoting the safe development of artificial intelligence.