Last year, the emergence of the AI language model ChatGPT sparked a global AI frenzy. With the release of the video generation model Sora, breakthroughs in AI technologies, represented by AIGC (Artificial Intelligence Generated Content), have significantly accelerated humanity's march into the age of intelligence. The development of large models and other AI technologies has become a core force leading the new generation of industrial revolution.
At this year's National Two Sessions, "AI+" was written into the government work report for the first time. "AI+", what should be added? How to add? What should be paid attention to? On March 24th, at the "AI+" Academician Forum held in Shenzhen, several academicians shared their views on "AI+", with "computing power" and "data" being the keywords frequently mentioned.
In the era of the digital economy, computational power is indispensable. Last October, the Ministry of Industry and Information Technology and five other departments jointly issued the "Action Plan for High-Quality Development of Computational Infrastructure," which states that by 2025, the scale of computational power will exceed 300 EFLOPS, with the proportion of intelligent computational power reaching 35%. In terms of storage capacity, by 2025, the total storage will exceed 1800 EB, with the proportion of advanced storage capacity reaching more than 30%.
Data shows that China's current operational data center rack total scale exceeds 7.6 million standard racks, with a total computational scale reaching 197 EFLOPS (1.97 x 10^17 floating-point operations per second), ranking second globally. This indicates that computational power, as a new type of information infrastructure, plays an extremely important role in the era of the digital economy. It's a key productive force and a "must-contest area" for the new round of technological revolution and industrial transformation.
"In the intelligent era, computational power will become a new type of productive force for social development, just like the transition from electricity to the power grid. The computational power network is an inevitable trend in the development of computational infrastructure," said Gao Wen, an academician of the Chinese Academy of Engineering and director of the Pengcheng Laboratory, at the forum. He mentioned that building China's computational power network still faces technical challenges in terms of core computational power supply, communication connections, and computational power scheduling.
In response, the Pengcheng Laboratory proposed the "China Computational Power Network" innovation plan. Based on the strong computational foundation of the Pengcheng Cloud Brain, the core hub of the "China Computational Power Network," the laboratory has trained the Pengcheng·BrainSea Large Model, the country's first fully independent, secure, open-source natural language pre-training large model foundation. It supports key domain model applications, aiding the development of a domestically innovative large model ecosystem.
"In recent years, artificial intelligence has made significant breakthroughs, thanks to big data, large models, and great computational power. These three elements are indispensable. However, simply increasing computational power is not a panacea. It's also important to consider how AI can serve the public interest," stated Li Guojie, an academician of the Chinese Academy of Engineering and a researcher at the Institute of Computing Technology of the Chinese Academy of Sciences, at the forum.
Shen Xiangyang, a foreign member of the National Academy of Engineering of the United States and a chair professor and president of The Chinese University of Hong Kong, Shenzhen, believes, "Large models generated by artificial intelligence technology will sweep through various vertical industries. Computational power is a key threshold, while data and talent are equally indispensable."
Shen mentioned that large models are a concept that integrates technology, market, governance, and other elements. They will profoundly affect all aspects of people's work and life, bringing new opportunities and challenges to the economy, society, and industries. Faced with the rapid development of large models, how to guide them with ethics, regulate them with laws, and bind them with morality, effectively unleashing the power of technological breakthroughs to achieve a balance between security and development, and integrating technology into society for the benefit of humanity, is an important question that must be answered well.
From text to text, images to videos, from AI-assisted surgical robots to AI teaching assistants, from making coffee to coding, from facial recognition payments to autonomous driving... The rapid iteration of artificial intelligence technology brings newer life experiences and production methods to the public. However, "Artificial Intelligence+" is not simply a matter of addition; it's about empowering various industries through new technology as a productive force.
According to Academician Li Guojie, artificial intelligence technology should consider how to serve the public interest more. He mentioned that China should combine its national conditions to carve out its own path of AI development, where AI technology should be applied more in fields such as materials, medical care, and industrial control to generate tangible economic benefits.
Wei Hua, an academician of the Chinese Academy of Engineering and a professor at Central South University, proposed two considerations: "The first is why we use large models and what industrial problems large models can solve. This must be clear. It's not about using large models for the sake of using them; there should be a goal orientation. The second is how industrial data and knowledge can be integrated with large models to produce tangible results."
"Artificial intelligence is not just about productivity issues; it's also necessary to consider its impact on the relations of production, social ideology, and optimization," believes Chen Chun, an academician of the Chinese Academy of Engineering and a professor at Zhejiang University.
When discussing "AI+ Education," Academician Lv Jian emphasized the importance of building knowledge systems, integrating knowledge systems, thinking systems, execution systems, and creation systems, and ensuring that what children learn can be applied to specific actions, achieving "unity of knowledge and action."
In the discussion session, attending academicians also delved into the future development of artificial intelligence technology, application scenarios, ethics, and other aspects. Everyone agreed that the development of artificial intelligence technology will bring significant changes to human society, but it's also necessary to seriously address the issues and challenges involved, ensuring that technological development truly benefits humanity.
