随着2025世界机器人大会和2025世界人形机器人运动会相继在北京召开，具身智能成为了人工智能与机器人领域的新热点。让我们一起来了解一下吧。
（一）具身智能
Embodied Intelligence
1. 什么是具身智能？
What is embodied Intelligence?
七十五年前，人工智能之父、英国计算机科学家艾伦·图灵（Alan Turing）就预见了机器智能发展的两个阶段：无实体智能与具身智能。当前的系统，如DeepSeek和ChatGPT，代表的是无实体智能；而具身智能则需要一个物理实体。
Seventy-five years ago, British computer scientist Alan Turing—widely considered the father of AI—foresaw two stages in the development of machine intelligence: disembodied intelligence and embodied intelligence. Current systems like DeepSeek or ChatGPT represent disembodied intelligence, while embodied intelligence necessitates a physical entity.
会话型机器人距离人工智能的终极目标仍有很长的路要走，而具身智能才是实现人工智能真正变革性影响所必需的。具身智能意味着智能是具身的、有情境的，它通过与现实世界的环境互动而发展，而非依赖于预先编程的知识或任务。近年来，神经网络、大模型和感知技术的突破，重新点燃了人们对具身智能这一概念的兴趣。
While a conversational machine is still far from being the ultimate goal of AI, embodied intelligence is necessary for AI to be truly transformative. That means intelligence is embodied and contextual, emerging through interaction with real-world environments rather than relying on pre-programmed knowledge or objectives. Recent breakthroughs in neural networks, large models and sensory technologies have reignited interest in the concept of embodied intelligence.
从机器狗、自动驾驶汽车到人形机器人，各类智能体正逐渐摆脱数据和屏幕的束缚，获得物理实体，迈入真实世界。
Intelligent agents, ranging from robotic dogs and autonomous vehicles to humanoid robots, are breaking free from the confines of data and screens. They are acquiring physical bodies and venturing into the real world.
2. 具身智能与传统人工智能的区别是什么？
What’s the difference between embodied intelligence and traditional AI?
具身人工智能的学习能力优于传统人工智能。传统的工业机器人需要工程师手动设定参数，而具身智能则通过三维视觉和模仿学习，就像人类一样，通过观察来掌握技能。这使得机器人能够像大脑与身体融为一体那样，模仿人类的动作。此外，这些机器人还具备泛化能力。在学会抓取一个杯子后，具身智能很可能无需重新编程就能抓取水瓶。
Embodied AI can learn better than traditional AI. Traditional industrial robots require engineers to program parameters, whereas embodied intelligence uses 3D vision and imitation to learn by watching—just like humans. It enables robots to mimic human movements as if the brain and body are integrated. Moreover, robots develop generalization capabilities—after learning to grasp a cup, they can likely grasp a water bottle without reprogramming.
目前，通用人工智能的应用主要集中在内容生成、客户服务、编程等桌面场景中，在工业领域的应用仍较为有限。这是因为工业环境是三维的，而大多数大模型是基于二维数据（包括语言、图像和视频）训练出来的。
Currently, general AI applications are concentrated in content generation, customer services, programming and other desktop scenarios, with limited industrial use. This is because industrial environments are 3D, while most large models are trained on 2D data, including language, images and video.
具身智能将人工智能深度融入三维物理环境之中，增强了其在真实场景中的感知、理解、交互与决策能力，并重新定义了人类与机器在物理世界和虚拟世界中的关系。它极大地拓展了人工智能的理解深度与广度，使其能够做出更科学、更理性、更具适应性的决策。
Embodied intelligence deeply integrates AI into 3D physical environments, enhancing perception, comprehension, interaction and decision-making in real-world contexts and redefining relationships between humans and machines in both the physical and virtual worlds. It vastly expands AI's depth and breadth of understanding, enabling more scientific, rational and adaptive decisions.
（二）人形机器人
Humanoid Robots
全球首届2025年世界人形机器人大会于8月17日在北京落下帷幕。该赛事吸引了来自16个国家的280支队伍，参与了包括竞技类、展示类以及场景应用类在内的26项比赛。这场金属、芯片与算法的较量，为以人形机器人为代表的具身智能新兴领域提供了一个面向公众的展示平台与技术交流的舞台。
The 2025 World Humanoid Robot Games, the first of its kind globally, which wrapped up on August 17 in Beijing, featured 280 teams from 16 countries competing across 26 events in categories such as competition, exhibition and scenario-based events. This clash of metal, chip and algorithm provided a public showcase and technical exchange platform for the emerging field of embodied intelligence represented by humanoid robots.
比赛项目包括100米、400米、1500米跑，立定跳远，自由体操和足球等。展示类项目则有人机舞蹈（包括单人舞与团体舞）、武术表演等，而场景类比赛则模拟了工厂、医院、酒店等真实环境。这些挑战共同构成了对人形机器人能力的终极考验。
The competition included events such as the 100m, 400m, 1500m, standing long jump, free gymnastics and soccer. The exhibition featured individual and group dance, as well as wushu, while scenario-based events simulated real-world settings such as factories, hospitals and hotels. Together, these challenges served as the ultimate test for humanoid robots.
在拳击比赛中，机器人的多个关节需高度协同，反应速度极快。当受到外力冲击时，机器人必须迅速恢复稳定与平衡。这类实践为开发者提升机器人性能提供了宝贵经验。而机器人足球则是技术最基础也最通用的测试场。人形机器人在足球场景中锤炼出的移动能力、感知算法、定位导航以及决策逻辑等能力，都具有在日常生活和专业场景中落地的潜力。
During boxing, multiple joints of robots need to coordinate with extremely fast reaction times. When subjected to external forces, the robots must quickly regain stability and balance. This practice provides valuable experience for developers to improve robotic performance. Robot soccer represents the most fundamental universal testing ground for technology. The capabilities honed by humanoid robots in soccer scenarios, such as mobility, perception algorithms, positioning and navigation, as well as decision-making logic, all hold potential for real-world applications in both daily life and professional scenarios.