Research on the Copyright Dilemma and Solution Path of Generative Artificial Intelligence Using Previous Works
The rapidly development and embedded applications of generative artificial intelligence pose challenges to the existing copyright system.Generative artificial intelligence normalizes the crawling of mas-sive amounts of prior work data,inducing infringement risks.Long distance text semantic understanding abil-ity is intended to cover up infringement traces,and open cross domain generalization reasoning ability pro-vides technical convenience for infringement.In terms of data feeding in works,generative artificial intelli-gence may cross the boundaries of fair use systems or lead to the extreme of unprotected or overprotected rights in previous works;In terms of copyright-ability of generated content,generative artificial intelligence decouples copyright subjectivity with its high-quality and massive generation ability,meeting the"minimum creative standards"and originality requirements.Propose to establish an endorsement system for the use of prior works to address the problem of rational use of algorithms;Promote the horizontal transition from"au-thor centrism"to"work centrism",and shift the authors rights law starting from personality rights to copy-right law starting from property rights;Establish an interpretable generative artificial intelligence originality evaluation mechanism and reinterpret originality standards.
Generative artificial intelligencePrior work dataCopyrightabilitySubjectivityOriginalityEndorsement mechanism
万方数据
维普
