脑机接口生成内容著作权挑战及其解决路径

Copyright Challenges and Solutions of Brain Computer Interface Generated Content

脑机接口是一种不依赖大脑正常输出通路在人与外部电子设备之间建立的全新交互系统,通过采集脑电波信号,转换为外部电子设备对大脑原始数据的处理符号呈现于显示屏上,供人类读取。本文通过研究脑机接口生成内容带来的著作权挑战,提出用户使用脑机接口生成内容,在消除神经元噪音与自主思维信号并增加制作成本以显示生成内容,进而达到独创性高度的智力劳动成果能受著作权保护。脑机接口智能体本身和脑机接口程序设计者不宜对脑机接口生成内容享有著作权,仅有投入智力劳动进行创作的用户能享有脑机接口生成内容著作权。脑机接口生成内容会产生两类著作权侵权情形:用户使用脑机接口生成与传统作品实质性相似的内容;脑机接口大模型训练使用在先传统作品或其他脑机接口生成内容。立法和司法实践可能支持这两类侵权的合理使用抗辩:用户使用脑机接口生成文字、音乐、美术等艺术作品可能落入我国立法条款列举的为个人、研究或欣赏以及适当引用合理使用情形;脑机接口大模型训练可能落入类似于谷歌图书扫描案的非表达转换性合理使用情形。

Brain computer interface(BCI)is a novel interactive system established between humans and external electronic devices not relying on the normal output pathways of the human brain.The BCI collects brainwave signals and converts them into symbols for processing raw brain data by external electronic devices,which are displayed on a screen for human reading.By studying the copyright challenges brought out by the BCI generated content,this article proposes that BCI generated content can be protected by copyright after eliminating neural noise and autonomous thinking signals as well as increasing production costs to guarantee that the BCI generated content is a highly original intellectual achievement.The BCI itself and the designer of the BCI program should not enjoy copyright over the generated content.Only users who have invested intellectual labor in creation can enjoy copyright.The generation of content through BCIs can result in two types of copyright infringement:users utilizing BCIs to generate content that is substantially similar to previous traditional works;BCI designers reproducing previous traditional works or BCI generated content to train large models.Legislations and judicial practices in China may support fair use defense for these two types of copyright infringement:users utilizing BCIs to generate text,music,art and other works may fall under the legislative exceptions of private study,research or appreciation,and appropriate citation;The training of BCI large models may fall into non-expressive transformative fair use offered in Google books copyright infringement case.