"Non-Work Use"Nature of Generative Artificial Intelligence Data Training and Its Legitimization
Whether or not the use of works in generative AI data training should fall within the scope of copyright rights is an im-portant stake in industrial development.The use of works in data training is"non-specific"and does not point to a specific indi-vidual work;from the perspective of functional stage,it belongs to the intermediate use in the production process,and from the perspective of subsequent effect,it has an enhancement effect on the basic tool of generative artificial intelligence.Therefore,it should be defined as"non-work use behavior"and excluded from the scope of copyright rights.Generative AI applications do not weaken the overall incentives for copyright holders,but only lead to a structural adjustment of incentives,and thus do not need to be compensated by extending copyright protection to data training.Of the specific options available under copyright law to exclude protection for"non-work uses"of training data,the"direct exclusion"from the scope of rights is generally superior to the fair use option.
generative artificial intelligencedata trainingcopyrightfair use
NETL
NSTL
万方数据
