生成式AI作品於著作權保護要件之挑戰

Regulatory Issues for the Application of Generative AI in Healthcare

由於深度學習技術的進展，人工智慧可取代人類創作內容來生成文字、影音、圖像等，未來的藝術、文學、或音樂等創作都可能是由人工智慧所產生，影響領域涵蓋生物醫療、工業、金融、或其他服務產業。美國著作權局公布之著作權與人工智慧報告認為僅在人類作者明確為充足的表達要件情形下，AI所產生的成果方能受到著作權保護。此涵蓋在AI生成過程中可察覺是由人類創作之情況，或人類作者對輸出之作品進行創意性安排或修改，而非僅給予AI提示。本文以生成式AI之作品是否符合著作權保護之要件為主要探討核心，有關生成式AI的保護仍應以人類精神創作投入為重點，單純由AI生成之作品不受保護。在構想與表達區分原則方面，特別針對運用生成式AI模仿風格之情形，應考量AI創作之迅速程度與便利性和傳統的人類作者之差異而進一步思考對原作者作品之保護是否充足。

As the advancement of deep learning technology, artificial intelligence can replace human creation to generate text, audio, video, images, etc. In the future, art, literature, or music creations may all be produced by artificial intelligence.“Copyright and Artificial Intelligence”published by the U.S. Copyright Office concludes that the results produced by AI can only be protected by copyright when the human author clearly provides sufficient expression requirements. This covers situations where the AI output process is detectable as a work created by humans, or where human authors make creative arrangements or modifications to the output work, rather than just giving the AI prompts. This article focuses on exploring whether the works of generative AI meet the requirements for copyright protection. The protection of generative AI should still focus on the creative input of human spirit. AI-generated works are not protected. Regarding the interpretation of the idea-expression distinction, especially for the use of generative AI to imitate artistic style, it should be considered that the speed and convenience of AI creation are still incomparable to traditional human authors. Therefore, further consideration can be given to how to provide more protection for the original creators.