生成式AI内容使用者的专利犯罪风险与刑法规制

Criminal Risks of Patent Infringement by Users of Generative AI Content and Criminal Law Regulation

摘要: 随着人工智能生成技术以及算法和大模型的更新迭代，人工智能生成软件(以下称AI软件)进入大众的生活，尤其是在工作中提供便利且可直接适用的方案、实验数据等。使用者信赖AI生成的内容，若直接使用于商业活动中，可能面临侵犯他人专利权内容产生的刑事法律风险。在AI的使用中就出现了AI的无心之失和使用者的有意之罪，传统专利法以及《刑法》第216条假冒专利罪，在应对目前隐蔽性高、极可能引发大规模侵权的新型风险可能存在处罚漏洞。本文通过分析AI内容使用者的行为模式和主观认识，论证如存在间接故意或重大监督过失，并造成严重后果的刑事可罚性进行探讨，以比较法学经验的路径提出针对这一新型风险的审慎刑法规制路径。

Abstract: With the iterative upgrading of artificial intelligence generation technology, algorithms and large models, artificial intelligence generation software (hereinafter referred to as AI software) has stepped into people’s daily lives, and in particular, it provides readily usable solutions, experimental data and other materials to facilitate work. When users trust and directly apply AI-generated content to commercial activities, they may face criminal legal risks arising from the infringement of others’ patent rights contained in such content. The use of AI has thus given rise to inadvertent errors made by AI itself and intentional crimes committed by its users. Traditional patent law, as well as the crime of patent counterfeiting stipulated in Article 216 of the Criminal Law, may have punishment loopholes in addressing such new risks that are highly concealed and likely to trigger large-scale infringements. By analyzing the behavioral patterns and subjective cognition of users of AI-generated content, this paper demonstrates the criminal punishability of such users when they act with indirect intent or gross supervisory negligence and cause serious consequences, and puts forward a prudent criminal law regulation path for this new type of risk by drawing on the experience of comparative law.