This paper examines the patent law issues of inventions utilizing artificial intelligence (AI) in the invention process, focusing on the criteria for inventive step and the disclosure requirements in the patent specification. Under current law, only h...
This paper examines the patent law issues of inventions utilizing artificial intelligence (AI) in the invention process, focusing on the criteria for inventive step and the disclosure requirements in the patent specification. Under current law, only humans are recognized as inventors. Therefore, if a human uses AI to complete an invention, the AI is not recognized as the inventor. However, if there is significant creative contribution from the human, the human is recognized as the inventor. If a human's creative contribution is recognized, the determination of inventive step should be based on whether a person skilled in the art could easily create the invention using AI. While maintaining the conventional “person having ordinary skill in the art” standard, the human's creative contribution should be considered. Furthermore, if the use of AI is omitted from the specification and the invention is presented as a human invention, this could constitute an unauthorized application. Therefore, the use of AI and the human's creative contribution should be specifically described. Furthermore, considering the “black box” nature of AI and the reproducibility issues, the specification should specifically disclose training data, algorithms, environments, and input values to meet the requirements of reproducibility and enablement.