基于ChatGPT的弱人工智能体犯罪刑法规制原则的确立与适用

Establishment and Application of Criminal Law Regulatory Principle of Weak AI Agent Crime Based on ChatGPT

目前阶段的ChatGPT仍属于弱人工智能体的范畴。涉弱人工智能体犯罪可能表现为侵犯公民人身权利、民主权利,侵犯财产权利,破坏社会主义市场经济秩序等诸多类型。刑法应确立安全与发展相平衡、刑事责任与法益侵害相关联、刑法规范完善与社会事实发展相同步三大规制原则。确立规制原则有利于人工智能刑法学的发展,对相关立法、司法实践也均有所裨益。将规制原则适用于实践中所带来的结果包括:应当追究过失犯罪的刑事责任;行业规范、研发生产者对产品的自我承诺可以作为研发生产者注意义务的来源,购销合同不能作为使用者注意义务的来源;宜以“行为当时的科技水平”作为预见可能性的判断标准;被允许的危险和企业合规对过失犯罪的刑事责任起到限制作用;对于现行刑法规定无法妥当定罪量刑的情形,应及时完善立法予以应对。

At the present,ChatGPT is still categories as Weak Artificial Intelligence(AI).Crimes involving Weak AI may be manifested as infringement of citizens' personal rights,democratic rights,infringement of property rights,destruction of socialist market economic order and so forth.As criminal law is concerned,three major regulatory principles should be established,that is,the balance between security and development,the correlation between criminal responsibility and infringement of legal interests,and the synchronization between the improvement of criminal law norms and the development of social facts.Establishing the regulatory principle is beneficial to the development of AI criminal jurisprudence,to the relevant legislation and to the judicial practice.The results of applying the regulation principle to practice include that the criminal responsibility of negligent crime should be investigated,and industry norms and self-commitment of R & D producers to products be the source of R & D producers' duty of care,whereas purchase and sales contracts cannot be the source of users' duty of care.It is advisable to take “the scientific and technological level of the behavior in the presence” as the judgment standard of the prediction possibility;permissible danger and corporate compliance limit criminal liability for negligent crimes.As for the situation that cannot be properly convicted and sentenced in accordance with the current criminal law,legislation should be perfected in time to deal with it.