Determining Copyright Ownership of AI-Generated Text-to-Image and Text-to-Video Works in the Context of Human-Machine Co-Creation

The widespread application of generative artificial intelligence in text-to-image and text-to-video scenarios has transformed creation into a human-machine collaborative process, challenging the traditional attribution of rights under copyright law. Current academic discourse lacks a comprehensive theoretical framework for determining copyright ownership of AI-generated content. A novel framework centered on the "operator as rights holder" offers a promising solution. From a legislative perspective, operators should be recognized as qualified rights holders when they make minimal yet objectively identifiable personalized choices during the creative process, thereby contributing originality. This approach to rights allocation aligns with the fundamental purpose of copyright law and reflects the technical reality that operators effectively control the means of expression. At the same time, operators should bear corresponding obligations to inform the public about the human-machine collaborative nature of their works. Copyright law may establish relevant certification and disclosure mechanisms through legislative and technological measures.