近日，众擎科技CEO赵同阳在一次行业峰会上分享了他对具身智能（Embodied Intelligence）的最新思考。赵同阳指出，具身智能不应仅被理解为机器人拥有感知与行动能力，而更应强调其在真实物理环境中通过持续交互学习、演化认知的能力。他认为，当前人工智能的发展正从‘数据驱动’迈向‘环境驱动’，具身智能将成为连接AI与现实世界的关键桥梁。赵同阳进一步强调，未来的人工智能系统需要具备‘身体’——无论是物理实体还是虚拟代理——才能真正理解人类语言、意图与社会规则。他以众擎正在研发的多模态智能体为例，说明如何通过模拟环境中的试错与反馈机制，使AI在动态场景中形成类人的推理与决策能力。他还呼吁产业界加强跨学科合作，融合认知科学、神经科学与机器人学，共同推动具身智能从实验室走向实际应用。赵同阳表示：‘真正的智能不是孤立运行的算法，而是在与世界的互动中不断成长的生命体。’这一观点为具身智能的发展提供了新的方向，也彰显了众擎在通用人工智能探索中的前瞻布局。
Recently, Zhao Tongyang, CEO of Zhongqing Technology, shared his latest insights on embodied intelligence at an industry summit. Zhao emphasized that embodied intelligence should not be narrowly viewed as robots merely possessing perception and action capabilities, but rather as systems that continuously learn and evolve cognition through interaction within real physical environments. He noted that AI development is transitioning from ‘data-driven’ to ‘environment-driven,’ with embodied intelligence serving as a crucial bridge between artificial intelligence and the real world.Zhao further stressed that future AI systems must possess a ‘body’—whether physical or virtual—to genuinely understand human language, intentions, and social norms. Citing Zhongqing’s ongoing development of multimodal agents, he illustrated how trial-and-error mechanisms and feedback loops in simulated environments enable AI to develop human-like reasoning and decision-making in dynamic scenarios.He also called for greater interdisciplinary collaboration across cognitive science, neuroscience, and robotics to accelerate the practical application of embodied intelligence beyond the lab. ‘True intelligence isn’t an isolated algorithm—it’s a living system that grows through continuous interaction with the world,’ Zhao stated. This perspective offers a fresh direction for the field and highlights Zhongqing’s forward-looking strategy in the pursuit of general artificial intelligence.
原创文章，作者：admin，如若转载，请注明出处：https://ktez.cn/5211.html