In recent times, many parties are accelerating the pace of innovation around artificial intelligence big models, conducting research on big model innovation algorithms and key technologies, accelerating the construction of intelligent computing infrastructure, and promoting the application of general artificial intelligence technology innovation scenarios.

Accelerate layout through multiple means

From the central to local levels, policy benefits are frequently released, accelerating the development of general artificial intelligence.

It was emphasized that we should attach importance to develop universal artificial intelligence, create an innovative ecosystem, and pay attention to risk prevention during a meeting of the Political Bureau of the Central Committee of the Communist Party of China,

Zheng Zhajie, Director of the National Development and Reform Commission, recently issued a document stating that we should grasp the direction of digitization, networking, and intelligence and emphasize the development of universal artificial intelligence.

The Several Measures to Promote the Innovative Development of General Artificial Intelligence in Beijing (2023-2025) (Draft for Soliciting Opinions) proposes 21 specific measures around five major directions, including "conducting research on big model innovation algorithms and key technologies" and "strengthening the research and development of big model training data collection and governance tools". At the same time, it expands application scenarios in fields such as government services, healthcare, scientific research, finance, autonomous driving, and urban governance to seize the development opportunities of large models and promote innovation and leadership in the field of general artificial intelligence.

Many domestic enterprises have also accelerated their business layout and launched their own artificial intelligence large model products.

An Hui, Deputy Chief Engineer of China Center for Information Industry Development, told Economic Information Daily that general artificial intelligence is considered a more advanced form of artificial intelligence technology and products. The emergence of large models reflects the trend of artificial intelligence towards more intelligent and universal development, indicating that there will be more applications of interdisciplinary and cross domain artificial intelligence, promoting the upgrading of economic and social intelligence and digital development.

Liu Cong, Dean of iFLYTEK Research and Development Institute Institute, stated that the application of artificial intelligence large models is expected to empower more practical scenarios in the socio-economic field. For example, in the field of smart education, the application of large models can assist in precise teaching and continuously upgrade language learning abilities. In terms of smart healthcare, it can assist grassroots doctors in improving their level of diagnosis, treatment, and prevention of diseases.

Improve the supply of computing resources

The demand for high-intensity computing in artificial intelligence is based on strong computing power, and currently, multiple parties are further strengthening their ability to coordinate and supply computing resources.

The Ministry of Science and Technology stated that it will promote the innovation of artificial intelligence models and algorithms for major scientific issues, develop a group of specialized platforms for artificial intelligence driven scientific research in typical scientific research fields, accelerate the construction of the national new generation artificial intelligence public computing power open innovation platform, and support the heterogeneous integration development of high-performance computing centers and intelligent computing centers.

Locally, Beijing proposes to efficiently promote the construction of new computing infrastructure to support the research and development of large-scale language models, multimodal large models, large-scale fine neural network simulation models, and brain inspired neural networks with hundreds of billions of parameter quantities. Shanghai has proposed to rely on the Shanghai Artificial Intelligence Public Computing Service Platform by the end of 2023 to connect and dispatch more than 4 computing infrastructure, with the ability to dispatch intelligent computing power to over 1,000 PFLOPS (FP16).

Telecom operators are also actively engaged in the construction and deployment of intelligent computing infrastructure. China Telecom Chairman Ke Ruiwen believes that the rapid development of large model algorithms in recent years is driving the development of artificial intelligence towards a more universal, precise, and inclusive direction, driving the upgrading of cloud service intelligence. The rapid evolution of large models has put forward higher requirements for computing power, data, and algorithms.

The reporter learned that China Telecom will further increase investment, integrate and gather innovative resources, continue to promote the upgrading of original technologies such as Tianyi Cloud operating system, storage engine, and distributed database, continue to create a universal visual model, build a unified and massive algorithm warehouse platform, and achieve one-stop scheduling of four-level AI algorithms and computing power. China Mobile will further promote the iterative upgrade of new information infrastructure, using mobile cloud as the unified entrance, providing convenient, fast, and ready to use computing services.

Balancing development and security

Emphasizing the development of artificial intelligence, while creating an innovative ecosystem, it is also necessary to pay attention to risk prevention.

The Management Measures for Generative Artificial Intelligence Services (Draft for Soliciting Opinions) released by Cyberspace Administration of China proposes that the state supports independent innovation, promotion and application, and international cooperation of basic technologies such as artificial intelligence algorithms and frameworks, and encourages the priority use of secure and trustworthy software, tools, computing, and data resources. At the same time, it is proposed that generative artificial intelligence products need to apply for a security assessment before providing services.

Qi Xiangdong, Chairman of Qi'anxin Technology Group, believes that artificial intelligence technology can greatly improve productivity, but it also brings new security risks. Enterprises should pay more attention to data security and privacy protection issues to ensure the security and reliability of their business.

“Through massive security knowledge and data, based on artificial intelligence big model related technologies, we can train security models, develop security products, and apply them to various aspects such as threat detection and vulnerability exploring. At the same time, relevant companies should do a good job in advance layout to strengthen the cultivation of network security talents." Qi Xiangdong said that currently, Qi'anxin has begun to delve into the security of artificial intelligence models, making certain breakthroughs in deep forgery, deep forgery detection and other technologies, and can accurately identify false images and videos generated by various artificial intelligence forgery technologies.

An Hui believes that for general artificial intelligence, we should not only recognize its important value, but also pay attention to and prevent risks, and use laws, policies, technologies and other means to address them.