The Decision of the Central Committee of the Communist Party of China on Further Comprehensively Deepening Reform and Promoting Chinese Modernization, adopted at the Third Plenary Session of the Twentieth Central Committee of the Communist Party of China, emphasizes the establishment of a safety monitoring system for artificial intelligence. As a strategic emerging technology, AI has been deeply integrated into people's production and life, which has greatly benefited the improvement of work efficiency, and has also continuously spawned new industries, new forms of business, and new business models, continuously injecting new kinetic energy into economic and social development. As an important part of artificial intelligence, the rapid development of generative artificial intelligence technology covers a variety of fields such as virtual reality, health care, creative industries, etc., bringing more innovation and value to various industries. However, at present, it is also necessary to pay close attention to the challenges of data security, misinformation and other risks that may be triggered by generative AI, and adhere to the equal importance of development and safety.
The so-called generative AI technology (AIGC) refers to a new generation of AI technology based on algorithms, models, rules, etc., to generate text, sound, pictures, videos and other content. Arithmetic, algorithms, and data are the three core elements of the development of this technology, and the combined effect of these elements makes it have a self-generating mode, which can realize the construction of a closed-loop learning system, and through the use of massive data and information, the elements of independent learning, which in turn generates original professional knowledge content or products and can be continuously updated and self-upgraded and iterative. Based on these characteristics, more and more industries are using it as a new element to improve productivity levels, creating exponential value and giving rise to a large number of industrial applications, with education, entertainment, business and other fields all being involved.2022 In late 2022, the launch of ChatGPT by US-based OpenAI signaled significant progress in the field of text generation by this technology. The World Intellectual Property Organization (WIPO) not long ago released the Generative Artificial Intelligence Patent Landscape Report, which shows that from 2014 to 2023, China's generative artificial intelligence patent applications exceeded 38,000, ranking first in the world, six times that of the second-place United States.
Generative AI is a double-edged sword, and while strengthening technological R&D and innovation, efforts should be made to improve its safety and controllability, effectively identify potential risks and loopholes, and form a response strategy.The Interim Measures for the Administration of Generative AI Services, which came into force on August 15, 2023, states that the relevant competent authorities shall carry out supervision and inspection of the generative AI services in accordance with their duties, and that the provider shall cooperate with them in accordance with the law, and shall require that the training data be provided to the relevant authorities in accordance with their duties. Cooperate with them, explain the source, scale, type, labeling rules, algorithmic mechanism mechanism, etc. of the training data as required, and provide the necessary technical, data and other support and assistance. Therefore, in order to further promote the development of generative AI in the future, we should start from the levels of network security, data security, personal information protection, etc., adopt an inclusive and prudent attitude, and set up “safety valves” from the technical, ethical, legal and other aspects, so as to prevent the risks of generative AI.
On the one hand, improve the effective definition of relevant laws and regulations, reasonable and effective supervision. Accelerate the legislative process, and carry out classification management and legislation for key issues in the development of generative AI, such as data, content safety and intellectual property rights, so as to effectively guide and coordinate the management of the design and evolution of generative AI through policy and legislation. Clearly define the regulatory bottom line, formulate rules and industry norms for the use of corresponding technologies, strengthen the supervision of generative AI enterprises, platforms, and various subjects using the technology in colleges and universities for misconduct, and promptly discover and deal with irregularities. Shift the “traditional after-the-fact supervision mode” to the “ex ante prevention and assessment mode” and the “during-the-fact control mode”, and formulate norms for ex ante assessment, risk early warning and during-the-fact control, so as to effectively reduce the occurrence of safety incidents and minimize the number of safety incidents. Formulate norms for prior assessment, risk warning and mid-control, so as to effectively reduce the occurrence of security incidents and minimize the negative impact of security incidents on enterprises and universities. Accelerate the formulation of corresponding rules for the use of technology and emergency measures, and strengthen the supervision and constraints on enterprises and platforms applying generative AI.
On the other hand, continuously enhance data security. Establish supervision and management of the entire life cycle of data, including the control of false and erroneous information entry at the data collection and input stage, the control and supervision of data labeling at the data processing stage, and data authenticity testing and verification at the data output stage (secondary collection and input stage). Data security is guaranteed through technical means for key areas such as medical care, finance, e-commerce, and colleges and universities. Timely detection and treatment of irregularities in data, safeguarding data and privacy security, preventing losses caused by misuse of data, and ensuring its compliance with laws and regulations and social morality. (Outlook news agency)