国际法视野下ChatGPT对数字安全的挑战及对策  

The Countermeasures and the Challenges of ChatGPT to Digital Security from the Perspective of International Law

作　　者：公惟韬 GONG Weitao

机构地区：[1]复旦大学,上海200433

出　　处：《国际商务研究》2025年第2期63-74,共12页International Business Research

基　　金：国家社会科学基金重大项目“构建人类命运共同体国际法治创新研究”(项目编号:18ZDA153);国家社会科学基金重点项目“‘人类命运共同体’国际法理论与实践研究”(项目编号:18AFX025)。

摘　　要：作为《全球安全倡议》的重要延伸,数字安全具有国家、社会和个人三重维度的内涵,目的是维护国家安全,弥合数字鸿沟以促进数字资源平等利用,并加强个人隐私保护。OpenAI公司发布的ChatGPT是生成式人工智能的标志性进展,但在算法、算力和数据层面,ChatGPT却分别存在威胁国家安全、数字资源利用不平等和泄露个人隐私的数字安全风险。目前,各国生成式人工智能发展水平不一,既有的国际法体系罕有对生成式人工智能的专门性规范,难以为相关技术水平落后的国家提供必要的保护。应完善数字安全国际法律制度,并从国家、社会和个人维度构建数字安全国际法体系,在算法、算力和数据层面完善对虚假信息的识别、保障数字弱势群体权益、落实“知情—同意”原则,平衡技术发展与安全,保护国家安全,弥合数字鸿沟,加强个人隐私保护,落实《全球安全倡议》,应对ChatGPT等生成式人工智能对数字安全的挑战。As an important extension of the Global Security Initiative,digital security has three dimensions including national,social,and individual.The purpose of digital security is to maintain national security,bridge the digital divide for equal use of digital resources,and strengthen the protection of personal privacy.ChatGPT released by OpenAI is a landmark for generative artificial intelligence.However,from the aspects of algorithm,computing power,and data,ChatGPT faces digital security risks that threaten national security,equal utilization of digital resources and personal privacy.At present,the development of generative artificial intelligence technology in various countries is different,and there are few special norms for generative artificial intelligence in the current international law system,which is difficult to provide necessary protection for countries which lack concerning technology.It is necessary to improve the relevant international law system of digital security and build an international law system of digital security from the national,social and personal dimensions.At the algorithm,computing power and data levels,measures should be taken including improving the identification of false information,safeguarding the rights of digital disadvantaged groups,and implementing the“informed consent”principle.This is to balance technological development with security,protect national security,bridge the digital divide,and enhance personal privacy protection,so as to implement the Global Security Initiative and addressing the security challenges posed by ChatGPT and other generative a rtificial intelligence technologies.

关 键 词：Chat GPT 数字安全 人工智能 风险 监管自治 

分 类 号：D996[政治法律—经济法学]