多维度解读与选择:人工智能算法知识产权保护路径探析

Multidimensional Interpretation and Choice:Exploring the Path of Intellectual Property Protection for Artificial Intelligence Algorithms

人工智能算法是人工智能快速进步的核心,虽然它的设计逻辑、运行过程、外观样态与传统计算机程序算法具有区别,却并未超出作为"知识产品"的本质。人工智能算法可在著作权法、专利法、反不正当竞争法中获得保护。其中,著作权法与专利法以保护人工智能算法表达为基本方式,商业秘密保护路径则以保护人工智能算法内核为根本诉求,三种规则利弊互现。人工智能算法的发展有赖于前人技术的公开,国际社会对人工智能算法的发展也提出了可解释性、透明性、可问责制的伦理要求。专利法的"公开发明理论"与人工智能算法的需求相契合,将人工智能算法作为专利法中的技术方案进行保护,在一定程度上能平衡技术权利人与使用人的利益,防止算法黑箱的产生,维护社会的道德与秩序。同时,审查机关应该对违背公共秩序、伦理道德的算法发明以及作为医疗方法的算法发明排除于专利客体之外。

Artificial intelligence(AI) algorithmis the core of the rapid progress of AI,its design logic,operation process and appearance are different from traditional computer program algorithm,but it does not go beyond the essence of "knowledge products".AI algorithms can be protected under copyright law,patent law and unfair competition law,with copyright law and patent law protecting the expression of AI algorithms as the basic approach and trade secret protection path protecting the core of AI algorithms.The above three rules have advantages and disadvantages respectively.The development of AI algorithms is dependent on the disclosure of previous technologies,and the international community has put forward ethical requirements for the development of AI algorithms in terms of interpretability,transparency and accountability.The "disclosure of invention theory" of patent law is in line with the needs of AI algorithms,and the protection of AI algorithms as technical solutions in patent law seems to balance the interests of technology rights holders and users,prevent the creation of algorithmic black boxes,and maintain social morality and order.At the same time,the algorithm invention that violates public order and ethics and morality,and the algorithm inventions as medical methods should be excluded from the patent object.