Generative artificial intelligence refers to the technical means and methods that use machine learning technology to allow computers to automatically generate data and information of different modalities. On April 11, 2023, the Cyberspace Administration of China (CAC) issued the Measures for the Administration of Generative AI Services (Draft for Comments) (hereinafter referred to as the Measures), which is an important exploration of generative AI governance in China. From the perspective of rule of law, generative AI should strengthen governance in terms of value norms, behavior regulation and process supervision.
Establish clear value norms to achieve technology for good
Technology itself has no value dimension, but people will give it a value dimension and make it form a value tendency. Generative AI can be abused, misused, or used for unethical purposes. As generative AI technology becomes more and more advanced, there are growing concerns that it will be used maliciously, undermining citizens' privacy and social security, and some even worry that it may lead to "human extinction". In order to ensure the sustainable and healthy development of generative AI technology and protect the rights of citizens and society, it is essential to uphold the principle of technology for good. The state shall establish clear values for technology research and development and use through legislation, stipulating that technology research and development and use should abide by public order and good customs and social morality, so as to lead the correct direction of technological development, make it conform to human interests and benefit human development. Of course, coercive measures taken through legislation should follow the strategy, as legislation that is too advanced may hinder technological development and innovation. However, the provisions of legislation at the value level can have stability and have a value level guidance for future problems.
It is necessary to clarify, concretize and institutionalize the value orientation of generative AI development through legislation, clarify the ethical guidelines and moral requirements of generative AI, and must not undermine social public order and endanger national and social security, so as to regulate and guide the development and application of generative AI technology. At the same time, it is necessary to step up the establishment of a generative AI research and development review and licensing system, requiring developers to conduct scientific, ethical and legal reviews before conducting generative AI technology research and development, and obtain corresponding licenses or approvals. The Measures issued by the Cyberspace Administration of China stipulate that generative AI content must be legal and compliant, algorithm discrimination is prohibited, unfair competition must not be implemented, false information must not be generated, and the legitimate rights and interests of others must not be infringed, which has an important guiding role in the development of generative AI technology at the value level.
Effectively regulate R&D and use through legislation
At present, output content such as ChatGPT and DALL-E is indistinguishable from human-generated content, and its quality and efficiency are much higher than the level of ordinary people, which will greatly improve the work efficiency and ability level of human beings and enrich human work and life scenarios. However, the negative impacts and potential threats are also highly unpredictable and uncertain. Some countries have begun to tighten legislation on generative AI to regulate its development and use. Strengthening the behavior regulation of generative AI through the rule of law is an important measure to ensure the reasonable, responsible and sustainable development of generative AI applications. It is necessary to establish legal norms and regulations for generative artificial intelligence, and clarify the norms and restrictions for the development and use of technology, so as to provide a clear legal basis for the research and development and use of generative artificial intelligence. At the same time, because the research and development and use of generative AI have strong global characteristics, it is necessary to strengthen international cooperation and exchanges, formulate relevant international standards and industry norms, and promote the international rule of law of generative AI governance. Countries can actively participate in the activities of relevant multilateral cooperation mechanisms, promote international cooperation governance, and formulate common standards and rules to promote and improve the legal regulation of generative AI on a global scale.
Generative AI often requires large amounts of data to train, and with the widespread use of generative AI, the collection, storage, and processing of personal data has become commonplace, so strengthening data management and privacy protections is critical. The state may strengthen legislation in the field of data, requiring the legality of pre-training data and optimized training data sources for generative AI products, stipulating rules for data collection, storage, use and sharing, ensuring the legality and security of data, and effectively protecting personal privacy and maintaining social security. The responsibility of generative AI application providers to protect users' rights and interests shall be clarified, including users' right to privacy, right to know, right to choose, etc., and may provide that users can revoke user data for generative AI applications at any time.
Establish a special agency for compliance and legality supervision
As generative AI becomes more and more seamlessly integrated into production and life, it is important to strengthen supervision, and it is necessary to establish an effective regulatory system. In particular, it is necessary to set up a special generative AI regulatory agency or set up a special department under the existing relevant institutions to be responsible for the supervision of generative AI, ensure that its development and use comply with laws, regulations and ethics, and prevent abuse and improper use. Laws and regulations should require generative AI technology providers to transparently disclose how their systems operate and what content they generate, and clarify the social and legal responsibilities of generative AI developers, suppliers and users, so as to ensure legal accountability in the event of errors, damage or infringement of generative AI systems.
Regulators should have sufficient authority and resources to monitor and inspect generative AI compliance and impose enforcement penalties for violations. This includes setting up independent regulatory bodies, establishing review procedures, developing compliance assessment requirements, and more. At the same time, it is necessary to strengthen social supervision through public education, improve the public's understanding and awareness of generative artificial intelligence technology, and encourage the public to participate in the regulation and decision-making process of generative artificial intelligence to ensure the public's right to know. The law could require generative AI developers and users to establish information exchange and communication mechanisms to enable the public to understand the potential impact and scope of application of generative AI. The use of generative artificial intelligence will bring a series of new problems, such as: ethical issues, property rights issues, liability issues, privacy issues, security issues, etc., which need to be dealt with at the level of the rule of law, which requires continuous exploration of effective rule of law means in practice, playing its role in benefiting society, and avoiding its harm to society.
Source: Beijing Daily | The author is Xiong Guangqing, professor of the Law School of the University of International Business and Economics and distinguished researcher of the National Academy of Opening to the Outside World
Process Editor: U016
![What are the security challenges posed by new technologies such as generative AI? The forum was held in Shanghai[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)