Currently, artificial intelligence (AI) has become a powerful engine driving the vigorous development of global technology and industry. General Secretary Xi Jinping pointed out: “Leading the transformation of research paradigms with artificial intelligence, accelerating breakthroughs in scientific and technological innovation across various fields.” The Fourth Plenary Session of the 20th Central Committee proposed to accelerate high-level technological self-reliance and self-improvement, leading the development of new productive forces. Guided by this goal, China is deeply implementing the ‘AI+’ initiative. As an emerging and future industry, the robotics sector must adapt to the trend, focus on enhancing productivity, and leverage AI+ to accelerate the transition of the robotics industry to new productive forces, becoming the ‘AI + Robotics Industry’, creating greater economic and social benefits for the national economy and people’s production and life.
Developing the robotics industry is not about replacing humans
We must understand that robots are meant to extend human capabilities to complete tasks, not to replace humans.
The origin and evolution of the human brain have undergone a long and complex process. Today, as Homo sapiens, our flourishing is attributed to chance and adaptability in the natural environment. Our ancestors overcame challenges that other hominid species could not. At that time, Homo sapiens were neither the strongest, nor the fastest, nor the most numerous (ants and krill outnumber humans by far). It can be said that it is our large and complex brains that enable us to adapt to and influence this planet.
The brain is currently the most complex collection of matter in the universe, and its evolution has made humans the most advanced animals in the biological world. Professor Yang Lequn, chief scientist at Meta and Turing Award winner, mentioned in a report, “A typical large language model is trained on about 10^14 bytes of information, which is nearly the total of all publicly available text on the internet. It would take a person hundreds of thousands of years to read all this material; it is an enormous amount of information.” However, relying solely on text training for large models is insufficient; we must continue to explore the mysteries of the brain that have evolved over hundreds of millions of years. The emergence of generative AI signifies a future restructuring of the deep economic structure. When planning the development of the robotics industry, we should deeply consider the relationship and transformation between humans and robots, including the redefinition of knowledge work and the restructuring of the labor force, rather than simply replacing humans on a large scale with robots. Therefore, we must focus on enhancing productivity, leveraging AI+ to accelerate the transition of the robotics industry to new productive forces, allowing robots to efficiently complete tasks as extensions of human capabilities. The result of robots will not be the end of jobs, but the reconstruction of work tasks.
We need to study the changes in manufacturing during different industrialization periods and the positioning of robots. In different industrialization periods, the positioning of factories in manufacturing is changing. Germany has proposed Industry 4.0 (the Fourth Industrial Revolution), while China is transitioning from traditional industrialization to new industrialization. With the advancement of technological changes led by artificial intelligence, most Chinese enterprises will evolve from automation to intelligence, and production models will shift from rigidity and standardization to flexibility and customization.
Robots will gradually transition from past automation tools to ‘AI + Robotics’; the controllers of robots will evolve from past real-time operating systems to ‘AI + Robotics’ intelligent systems; automation robots interacted through buttons, keyboards, mice, and screens, while ‘AI + Robotics’ can enhance voice interaction; automation robots were programmed manually, while ‘AI + Robotics’ relies on large models; automation robots worked at fixed stations, while ‘AI + Robotics’ will achieve autonomous movement across multiple stations; automation robots required prior deployment, adjustment, and programming, while ‘AI + Robotics’ will enable plug-and-play; automation robots operated in a human-machine division of labor with limited interaction, while ‘AI + Robotics’ will enable human-machine collaboration, complementing each other’s strengths. In summary, ‘AI + Robotics’ is the future direction of robot development, and the shapes of robots will diversify, potentially completing specific tasks in the most economical and reasonable way according to scene requirements.
To effectively utilize robots, we must focus on three core intelligent collaborations
From the current industrial status, the key to the development of China’s robotics industry is to enhance the intelligence level of robots. The control of robots needs to evolve from the past ‘robot operating system’ to the ‘robot intelligent system’. It is generally believed that three core capabilities support the robot intelligent system: first, the ‘eyes’, which refer to the robot’s environmental perception capability; second, the ‘action’, which refers to motion control capability; and third, the ‘brain’, which refers to interactive decision-making capability. These three aspects are interrelated yet relatively independent, collaboratively forming a robot intelligent system. Currently, China’s robotics industry invests heavily in motion control (‘action’), while investment in environmental perception (‘eyes’) and interactive decision-making (‘brain’) is insufficient and urgently needs enhancement.
The large language model constitutes the ‘brain of the robot’. The robot intelligent system, centered around the ‘brain’, has driven a technological architectural transformation from ‘robot operating systems’ to ‘robot intelligent systems’ based on large language models. The large language model brings efficiency improvements and redesigns work processes. With the support of large models, robots can autonomously implement knowledge accumulation, accept instructions, execute actions, and engage in human-robot interaction, among other tasks. The large model will provide corresponding work processes to guide robots in completing various tasks.
AI + spatial computing serves as the ‘eyes of the robot’, opening a new paradigm for robots to understand the world. According to the principles of biological intelligence evolution, the eyes are the starting point of intelligence in biological evolution. The robot intelligent system must emphasize the role of ‘eyes’, and using AI + spatial computing as the support for the robot’s eyes is appropriate, allowing for learning and training of the environment in a ‘human-like eye’ manner using ordinary monocular cameras and neural network learning, enhancing perception and understanding of the physical world, and possessing adaptive and continuous learning capabilities, enabling robots to open their eyes to see the world, characterized by usability, ease of use, and practicality.
Spatial computing, as a computing model oriented towards the three-dimensional world, is reshaping the interaction between humans, machines, and the world, and is one of the key core technologies driving the implementation of robotics. Historically, the paradigm of human-machine interaction has evolved from telegraphy and text to voice, graphics, and video. Currently, we are transitioning from two-dimensional to three-dimensional new interaction paradigms, with spatial computing being an important technological support leading this transformation. Spatial computing is a ‘reconstruction’ of the physical world, which generative AI cannot replace. Today, the rapid development of generative large language models does not cover all aspects of our world. The data from the physical world is complex and diverse in modalities, and current large models cannot easily parse video, actions, physical rules, and so on.
The integration and innovation of AI + spatial computing are expected to break the boundaries between the physical and digital worlds, leading us into a new information age, and are key technologies supporting the development of low-altitude economies, robotics, and other industries. Its difference from traditional machine vision lies in the fact that it reconstructs the physical world using AI + spatial computing, while the latter often relies on the superposition of various physical hardware to perceive the world.
The open-source AGIROS supports ‘robot actions’ and contributes to the ecological construction of the robotics industry. AGIROS is an open-source robot operating system supported by the Software Institute of the Chinese Academy of Sciences, proposed as a standard. Its open-source community was initiated by the Software Institute of the Chinese Academy of Sciences, aiming to gather the strengths of various parties in the robotics industry, academia, and application to comprehensively promote open-source and collaborative innovation in the field of intelligent robotics, laying a solid foundation for the intelligent robotics industry. To date, a large number of units and developers in the robotics field have joined this community.
We believe that with the power of open source, AGIROS will greatly enhance the competitiveness of the ‘brain, eyes, and actions’ collaborative system of AI + robotics, leading the trend in the global robotics industry and becoming a major driving force for the leap from traditional robotics to AI + robotics.
Building an ‘AI + Robotics’ Ecosystem Based on RISC-V Architecture
Historically, there have been several large-scale ecosystems in the field of information technology, such as ‘Wintel’ and ‘AA’, and the emerging ‘RV + OSS’, including ‘RV + OpenHarmony’ and ‘RV + openKylin’, etc., all of which are information technology ecosystems oriented towards humanity. In the future, a similar ecosystem may also form for ‘AI + Robotics’, such as the ecosystem discussed in this article based on ‘AI + Robotics’ with RISC-V architecture (i.e., RV chips + brain, eyes, and action intelligent systems).
We hope that the Chinese robotics industry will make greater contributions to this ecosystem, and that we will work together to build a world of human-robot integration, allowing robots to become extensions of our capabilities and help humanity achieve the beautiful life we aspire to. Let us help robots build their own intelligent systems to see the world, understand the world, and act in the world.
Author Biography
Ni Guangnan, born in 1939, from Zhenhai, Zhejiang, is a researcher at the Institute of Computing Technology, Chinese Academy of Sciences. He graduated from Nanjing Institute of Technology (now Southeast University) in 1961, pioneering the application of associative functions in Chinese character input, and served as the first chief engineer of the company (the predecessor of Lenovo) and Lenovo Group, leading the development of the Lenovo Chinese character system and the Lenovo series of microcomputers, which won the National Science and Technology Progress Award in 1988 and 1992, respectively. Lenovo Group was founded based on the Lenovo Chinese character system and derived its name from it. Since then, he has been committed to developing autonomous and controllable core information technologies and industries. In 1994, he was selected as one of the first academicians of the Chinese Academy of Engineering, and in 2011 and 2015, he received the Lifetime Achievement Award from the China Society for Information Science and the China Computer Federation, respectively.
Source: Global Network
Recommended Reading
◆ How Many Academicians in Guangdong | Guangdong Academician MapCollection
◆ Symposium on Learning and Implementing the Spirit of the 2025 Provincial High-Quality Development Conference Held by Academicians and Experts in the Greater Bay Area
◆ Gathering Global Talent to Strengthen Guangdong Industry Towards New Forces! The 2024 Guangdong-Hong Kong-Macao Academician Summit and the 10th Anniversary Academic Annual Meeting of the Guangdong Academician Association Concluded Successfully◆ Nine Academicians Discuss Future Industry Frontier Technology! The 2024 Guangdong-Hong Kong-Macao Academician Summit and the 10th Anniversary Academic Annual Meeting of the Guangdong Academician Association Successfully Held◆ Innovating with Focus, Promoting Mutual Strength in Industry and Technology! The 10th Anniversary Cooperation Development Conference of the Guangdong Academician Association Held in Guangzhou