生成式人工智能反垄断论纲  

Outline of Anti-Monopoly for Generative Artificial Intelligence

作　　者：王健[1] 吴宗泽[2] WANG Jian;WU Zongze

机构地区：[1]浙江理工大学法政学院 [2]清华大学人工智能国际治理研究院

出　　处：《法治研究》2024年第6期130-147,共18页Research on Rule of Law

基　　金：国家社会科学基金重点项目“我国反垄断执法现代化的理论阐释与实现路径”(项目编号:24AFX016)阶段性成果。

摘　　要：生成式人工智能技术的发展,依赖于对数据、模型、人才和算力等关键资源的利用。通过分析各个关键资源背后的市场特征,可以发现其中有着内在的结构性垄断担忧。如果缺乏及时、有效的反垄断监管,生成式人工智能技术的发展易受到垄断行为的阻碍。依据垄断行为影响程度的强弱,可以将潜在垄断行为分为封锁型行为、限制型行为和剥削型行为三类。为了营造有利于技术创新的公平竞争市场环境,防止结构性的垄断担忧变成实质性的发展障碍,我国反垄断执法机构应当基于包容审慎的监管原则推进反垄断监管。在此基础上,采用事前预防性监管为主的监管策略,并且强化面向国际、产业的监管合作。The development of generative artificial intelligence(Generative AI)technology hinges on the utilization of key resources such as data,models,talent,and computing power.By analyzing the market characteristics behind each key resource,we can notice inherent structural monopoly concerns.Without timely and effective anti-monopoly supervision,the development of Generative AI technology can easily be impeded by monopolistic behaviors.Based on the intensity of the impact of monopolistic behavior,potential monopolistic behavior can be categorized into blockade,restrictive,and exploitative behaviors.To foster a fair and competitive market environment advantageous to technological innovation,and to prevent structural monopoly concerns from becoming substantial development obstacles,anti-monopoly law enforcement agencies should promote anti-monopoly supervision grounded in principles of inclusive prudence.Building upon this,a regulatory strategy that emphasizes proactive and preventative supervision should be adopted,along with strengthening regulatory cooperation at international and industry levels.

关 键 词：生成式人工智能 结构性垄断 垄断行为 反垄断监管 

分 类 号：D922.294[政治法律—经济法学] D922.17[政治法律—法学]