智源研究院发布2026十大AI技术趋势报告

Core Insights - The report from Beijing Zhiyuan Artificial Intelligence Research Institute highlights a significant shift in AI evolution from parameter scale in language learning to a profound understanding and modeling of the physical world [1][2] - The transition marks a move from functional imitation to understanding the laws of the physical world, indicating a clearer development path for AI that integrates into the real world to address systemic challenges [1] Group 1: Key Trends in AI - The competition in foundational models has shifted focus from "how large the parameters are" to "whether it can understand how the world operates" [1] - The new paradigm represented by "Next-State Prediction" (NSP) is pushing AI from "perception" in the digital space to "cognition" and "planning" in the physical world [1] Group 2: Driving Forces Behind the Transition - The transition to 2026 is driven by three clear main lines: the "upgrading" of cognitive paradigms, where AI begins to learn physical laws, providing a new cognitive foundation for complex tasks like autonomous driving simulation and robot training [2] - The "embodiment" and "socialization" of intelligent forms, where intelligence is moving from software to physical entities, with humanoid robots entering real production scenarios [2] - The "dual-track application" for value realization, where a super application portal is forming on the consumer side, and on the enterprise side, AI is yielding measurable commercial value products in vertical fields after the initial concept validation phase [2]