生成式人工智能的风险及其包容性法律治理

    Risks of Generative Artificial Intelligence and Its Inclusive Legal Governance

    • 摘要: 以ChatGPT为代表的生成式人工智能的诞生,标志着人工智能技术从“决策式AI”向“生成式AI”的重大技术跃迁,引发了从“时空革命”到“知识革命”的技术演进。生成式人工智能技术在带来众多科技红利的同时,也引发包括数据安全问题、个人隐私侵犯、知识产权侵害、误导性和虚假信息泛滥、市场垄断、偏见和歧视等诸多现实风险。为了鼓励生成式人工智能技术的创新发展以及规避其所带来的风险,中国采取“包容性法律治理”模式。在治理原则上,包容性法律治理始终坚持协同共治、提升透明度、提高数据质量、发挥法治效能等基本理念;在治理主体上,包容性法律治理提倡多方主体协同治理,充分发挥政府机构、科研机构、行业协会和公民团体等社会组织的力量;在治理过程中,包容性法律治理主张运用多种治理工具,通过技术方案、伦理标准和法律政策全面提升生成式人工智能的风险治理水平;在治理方法上,采取分类分级治理策略;在治理结果上,包容性法律治理强调要注重数字社会的分配公平,弥合数字鸿沟,保障数字空间的基本人权。

       

      Abstract: The birth of generative AI represented by ChatGPT marks a major technological leap from “decision-making AI” to “generative AI”, triggering a technological evolution from “space-time revolution” to “knowledge revolution”. It has triggered the technological evolution from the “space-time revolution” to the “knowledge revolution”. While generative AI technology brings many technological dividends, it also raises many real risks, including data security issues, personal privacy infringement, infringement on intellectual property rights, proliferation of misleading and false information, market monopolization, bias and discrimination. In order to encourage the innovative development of generative AI technologies and avoid the risks they pose, China has adopted an “inclusive legal governance” model. In terms of governance principles, inclusive legal governance always adheres to the basic concepts of collaborative governance, enhancing transparency, improving data quality, and giving full play to the effectiveness of the rule of law; in terms of governance subjects, inclusive legal governance advocates the collaborative governance of multiple subjects, and gives full play to the power of government agencies, scientific research institutes, industry associations, civic groups, and other social organizations; in terms of governance process, inclusive legal governance advocates the use of a variety of governance tools. In the governance process, inclusive legal governance advocates the use of multiple governance tools to comprehensively improve the risk governance level of generative AI through technical solutions, ethical standards and legal policies; in governance methodology, it adopts a classification and grading governance strategy; and in governance outcome, inclusive legal governance emphasizes the need to focus on distributional fairness in the digital society, to bridge the digital divide, and to safeguard the basic human rights in the digital space.

       

    /

    返回文章
    返回
    Baidu
    map