Transparency obligations in AI training data and the protection of intellectual property in Vietnam

In the context of rapid digital transformation, artificial intelligence (AI) systems are increasingly developed on the basis of heterogeneous training data, much of which comprises works and materials protected by intellectual property rights. The use of such data raises pressing concerns regarding the transparency and traceability of AI training sources, both to safeguard the lawful rights and interests of relevant stakeholders and to promote the responsible development of AI technologies. Through analytical, synthetic, and comparative examination of the 2025 Artificial Intelligence Law, the 2025 amended Law on Intellectual Property, and related legal instruments, this article finds that Vietnamese law has yet to establish specific obligations governing the disclosure or traceability of AI training data. This regulatory lacuna risks undermining the effectiveness of intellectual property protection and generating legal uncertainties in the deployment and application of AI systems. On this basis, the article advances several directions for reform, emphasizing enhanced transparency requirements for AI training data in order to foster a safe, transparent, and sustainable legal framework for scientific and technological development in Vietnam in the digital era.