Data mining or machine learning without permission may infringe upon multiple rights of copyright owners,such as right of reproduction,right to prepare derivative works,and right of information network dissemination.Although there are currently some internet platforms and databases available for data mining and machine learning,enabling users to obtain works in bulk,the failure of data miners and machine learners to seek prior permission still may lead to prohibitive transaction costs,hindering effective access to necessary data and constraining the development of digital innovation products.Allowing data miners and machine learners to freely use copyrighted materials may substantially harm copyright owners'market interests.However,the digital innovation products developed as a result may bring additional market benefits to copyright holders.Moreover,the self-regulating mechanism of society can mitigate many market harms.Therefore,disruptive improvements in machine learning and data mining that challenge existing technologies should be recognized as fair use.It is not advisable to regulate marching learning and data mining behaviors through statutory license or"opt-out"mechanisms for fair use.
万方数据
维普
