Transparency obligations are central to artificial intelligence (AI) regulation. The Framework Act on AI and its Enforcement Decree outline specific provisions regarding these obligations. While stipulating transparency obligations in legislation can promote legal stability, it also represents limitations in clearly defining applicable standards. Critics have highlighted the ambiguity of transparency obligations under the current legal framework, noting that legal uncertainty will remain difficult to eliminate entirely, even if the relevant provisions are supplemented in the future. Furthermore, stipulating transparency obligations may hinder flexible responses to emerging technologies and unpredictable future scenarios. Crucially, the current mandatory-centric approach stems from a negative perception of AI technology and may be inconsistent with the promotion of AI development and utilization. Therefore, the approach to AI transparency obligations should shift from mandatory regulation toward minimum safeguard framework that enable AI service providers to mitigate liability. Additionally, aligning with legislative trends in major countries, the scope of exemption through voluntary guidelines should be expanded.