Core Insights - The 2025 China Automation Conference (CAC2025) was held in Harbin, focusing on the integration of automation and intelligence, highlighting new opportunities for automation technology in conjunction with artificial intelligence and other advanced information technologies [1][2] - The conference serves as a platform for summarizing innovations in the automation field during the 14th Five-Year Plan and for discussing technological breakthroughs and industrial layouts for the upcoming 15th Five-Year Plan [1] Group 1 - Automation technology is seen as a core engine driving the advancement of intelligent manufacturing, with autonomous intelligent automation expected to be the backbone of future factories [1][2] - Autonomous intelligent automation aims to integrate advanced technologies such as AI, IoT, and big data, enabling systems to have real-time reconfiguration capabilities and to autonomously perceive, analyze, decide, and execute tasks [1][2] - In process industries, these systems can capture thousands of production parameter fluctuations in real-time, autonomously adjusting processes to stabilize product quality while reducing raw material consumption [1] Group 2 - Four key directions for the future development of automation systems were proposed: integrating first principles into manufacturing world models, achieving efficient data generation for industrial embodied intelligence, establishing a solid digital foundation for virtual controllers, and adapting manufacturing elements at multiple levels for industrial operating systems [2] - By 2040, advancements in autonomous intelligent technology are expected to lead to super automation systems in factories, where systems can operate autonomously based on expressed human intentions [2] Group 3 - Artificial intelligence is playing a dual role as both a "disruptor" and an "efficiency enhancer" in the transformation of industrial control systems, breaking the limitations of traditional control systems that rely on pre-programmed algorithms and human experience [3] - The integration of knowledge-driven mechanistic models with data-driven AI models creates a closed-loop system for perception, decision-making, execution, and feedback, allowing for rapid adaptation to changes in products, processes, raw materials, and environments [3] - The use of multimodal large models for automatic conversion of process design drawings to control codes has significantly reduced implementation costs and time for industrial control systems, while natural language human-machine interaction has lowered user training costs and operational difficulties [3]