摘要
既有防沉迷规范的适用范围有限并且没有指出相关主体的具体义务,其难以应对生成式人工智能令人沉迷的风险。从原因视角分析,生成式人工智能令人沉迷的原因在于其生成的虚拟世界更具有吸引力、精准推送产生舒适性、虚拟人物使人产生情感依赖。从后果视角分析,人们对生成式人工智能的沉迷容易造成创新能力退化、道德标准僵化与价值导向机械化、审美形式化等问题。基于此,生成式人工智能提供者和开发者首先需要履行审批和披露义务。其次应设置防沉迷层面的算法,包括根据用户的年龄等因素设置相应的防沉迷策略,设置用户偏好推送的限度,履行适当的披露义务并拒绝提供情感慰藉;应当对生成式人工智能在思维认知领域“作为思想生成、创新工具”加以限制;设置道德过滤机制;在特定情形中限制“直接提供成品”的使用频次。再次,定期进行算法检查。最后,生成式人工智能提供者和开发者违反防沉迷、防止依赖义务应当承担行政责任。
The existing anti-addiction norms are limited in scope of application and do not point out the specific obligations of relevant entities,which is difficult to deal with the risk of addiction in generative artificial intelligence.From the perspective of reasons,the reason why generative artificial intelligence is addictive is that the virtual world generated by it is more attractive,accurate pushing generates comfort,and virtual characters make people emotionally dependent.From the perspective of consequences,people's addiction to generative artificial intelligence tends to lead to the degradation of innovation ability,rigid moral standards,value-oriented mechanization,and aesthetic formalization.Based on this,generative artificial intelligence providers and developers first need to fulfill their approval and disclosure obligations;According to the user's age and other factors,set the corresponding anti-addiction strategy,set the limit of users preference pushing notification,fulfill the appropriate disclosure obligation and refuse to provide emotional comfort;Generative artificial intelligence should be limited in the field of thinking and cognition as a tool of thought generation and innovation.Set up moral filtering mechanism;Limit the frequency of use of“direct delivery of finished products”in certain circumstances;And the algorithm is checked regularly.Generative artificial intelligence providers and developers should bear administrative responsibility for violating the obligation to prevent addiction and dependence.
作者
李岩
康铭
LI Yan;KANG Ming(School of Law,Liaoning University,Shenyang 110036,China)
出处
《东北师大学报（哲学社会科学版）》
CSSCI
北大核心
2024年第2期112-126,共15页
Journal of Northeast Normal University(Philosophy and Social Science Edition)
基金
国家社会科学基金项目(22BFX181)
辽宁省“兴辽英才计划”项目(XLYC2004005)
辽宁省教育厅2021年度科学研究经费项目(LJKR0049)
辽宁省社会科学规划基金重大项目(L23ZD016)。
关键词
生成式人工智能
防沉迷义务
行政责任
Generative Artificial Intelligence
Anti-addiction Obligation
Administrative Responsibility
作者简介
李岩,男,辽宁大学法学院教授、博士生导师;康铭,男,辽宁大学法学院博士研究生。
