| Summary: | 碩士 === 世新大學 === 智慧財產權研究所(含碩專班) === 107 === The technology of artificial intelligence has achieved remarkable development in recent years, especially in the field of deep learning technology, which makes the improvement of artificial intelligence in creativity keep pace with human creation. The whole world is transforming with the fourth industrial revolution and the level of success of the industry will depend on the application of artificial intelligence technology. In other words, the fourth industrial revolution is providing a good environment for the development of artificial intelligence, and the breakthrough in deep learning technology provides the potential for its application in industrial production.
With the improvement of creativity of artificial intelligence however, a problem that was difficult to imagine in the past is now present today: infringement. For example, the reason about the creativity of artificial intelligence can be promoted to the same level as human beings is because of the improvement in machine learning. This has been a result of the breakthrough of deep learning technology in recent years. Under the operation of this mechanism, artificial intelligence can pave the way of learning, collecting many expressions that are difficult to detect in people's observation, which allows it to reproduce the essence of originality in a non-traditional way. However, if there is no clear authorization channel between training data and original work, then the creative process of artificial intelligence would infringe on intellectual property rights. This is especially true with the application of deep learning technology that artificial intelligence uses. Because humans have a considerable lack in understanding of its operational procedure, it is difficult for people to know which circumstances would constitute similarity between machine output and original work. In addition, the expression of artificial intelligence is more complicated than in the past. For example, artificial intelligence can add its own understanding of original work in the way of human creators, or reproduce the original work’s features in a concealed way. As a result, the standard of the similarity needs to be determined through further analysis. But more importantly, since artificial intelligence uses deep learning as a way to create work, what has now emerged is the debate of determining the independent status of artificial intelligence. Especially in the future, human beings will have difficulty grasping artificial intelligence creative mode. In the event of infringement, who bears the liability? This paper will focus the scope of discussion strictly on copyright issues.
In order to answer the above questions, I will focus on the analysis of technology and industrial application of artificial intelligence. I will explore what stage of development the creativity of artificial intelligence surpasses the point of human beings’ grasp of it. This will be supported by the national policy report and the research on the three artificial intelligence development eras. The issue of artificial intelligence creation infringement still belongs to the hypothetical issues discussed by academia. This article will use existing artificial intelligence creation instances and see how they apply to the current rule of fair use and copyright infringement. I will study to see if the creation of the artificial intelligence was the result of infringement and establish boundaries between infringement and fair use. In addition, based on the international concern about the application of the liability subject system of AI infringement, this paper will also explore the most appropriate liability for infringement by assuming that AI independently creates infringement. Finally, this paper will use the findings of this discussion as a reference for the artificial intelligence legislation policy. By doing this, our country will have the ability to prevent confusion in the future in AI creation debates.
|
|---|
