人工智能学习创作的侵权困境与出路

Infringement Dilemma and Outlet of Artificial Intelligence Learning Creation

近年来,以大数据、人工神经网络和深度学习为基础的人工智能技术迅猛发展,并在文学艺术领域涌现出极具欣赏价值的人工智能创作物。但是,研发人员在训练人工智能学习创作的过程中,往往需要使用他人作品数据,而这一行为可能会侵犯作品著作权人的权利。为平衡作品著作人的利益以及人工智能技术发展的公共利益,在立法层面,应对著作权法定许可制度和集体管理制度进行完善,以对人工智能使用海量作品的行为予以规制,进而在保障相关权利人合法利益的同时,促进人工智能技术的有序发展。

In recent years, big data, artificial neural networks and deep learning have developed rapidly, and in the field of literature and art, there are so many AI great creations. However, in the process of training artificial intelligence learning, researchers often use the data of others’ works, which may infringe the rights of the owners.In order to balance the interests of the authors of works and the public interests of the development of artificial intelligence technology, at the legislative level, we should improve the statutory licensing system and collective management system of copyright, so as to regulate the use of large quantities of works by artificial intelligence and promote the orderly development of artificial intelligence while safeguarding the legitimate interests of relevant obligees.