摘要
随着以体育康复、技战术分析和体育文献为代表的体育大模型兴起,训练数据集作为其发展的核心要素,正面临着高质量数据短缺的困境。构建既能促进数据集供给又能保障数据相关权益的促进型数据制度,已成为亟待回应的关键议题。学界所提的“个人信息合规放宽论”“促进型规则供给论”“著作权制度调试论”,要么忽视从既有法律规定中探寻应对之策,要么对数据标注的促进型规范供给不够重视,以至于在一定程度上与实践需求相脱节。对此,需从法理层面剖析个人信息保护与模型发展之间的潜在冲突,揭示促进型数据规则赋能模型发展的限度,阐明著作权制度约束模型发展的阶段差异。以此为指导,体育大模型促进型数据制度宜从以下三方面进行体系优化:其一,以信息披露和影响评估为重点,健全个人信息保护重点制度的规范体系;其二,针对数商培育、数据集开源以及公共数据开放,制定“需求导向”的促进型规则;其三,细化不同阶段模型所涉主体的著作权责任划分,对同一主体在不同阶段的行为配置相应的义务。
With the rise of large-scale sports models represented by sports rehabilitation,technical-tactical analysis,and sports literature,the training datasets,as a core element of their development,are facing a shortage of high-quality data.Consequently,establishing a facilitative data governance framework that both promotes dataset supply and safeguards data-related rights has become a critical issue to be addressed.Academic proposals such as the“relaxation of personal information compliance theory,”the“facilitative rule supply theory,”and the“adjustment of copyright system theory”either overlook exploring solutions within existing legal frameworks or fail to sufficiently emphasize the supply of facilitative norms for data annotation,leading to a certain disconnection from practical needs.To address this,it is necessary to analyze the potential conflicts between personal information protection and model development from a legal perspective so as to reveal the limitations of facilitative data rules in empowering model development,and to clarify the phased differences in how the copyright system constrains model development.Guided by this,the facilitative data governance framework for large-scale sports models should be optimized in the following three aspects.Firstly,information disclosure and impact assessment should be emphasized in order to improve the normative system of key personal information protection institutions.Secondly,“demand-oriented”facilitative rules should be formulated for the cultivation of data businesses,dataset open-sourcing,and public data openness.Thirdly,division of copyright responsibilities among entities involved in different stages of model development should be refined and corresponding obligations to the same entity at different stages should be assigned.
作者
黎浩田
LI Haotian(School of Law,Wuhan University,Wuhan 430064,China)
出处
《武汉体育学院学报》
北大核心
2025年第3期46-52,102,共8页
Journal of Wuhan Sports University
基金
国家社会科学基金一般项目(24BFX052)
湖北省体育局决策咨询研究项目(2024B002)。
关键词
体育大模型
个人信息保护
著作权
促进型规则
large-scale sports model
personal information protection
copyright
facilitative rule
