데이터 페미니즘 관점에서 본 인공지능기본법

Data feminism challenges the assumption that data is neutral and objective by revealing how technological development reproduces gender inequality and entrenched power asymmetries. It advances both theoretical and practical strategies to address these structural injustices. To overcome gender inequality in data science, data feminism articulates core principles, including examining and challenging power, elevating emotion and embodiment, rethinking binaries and hierarchies, embracing pluralism, considering context, and making labor visible. From this perspective, Korea’s Framework Act on Artificial Intelligence fails to place gender equality at the core of its regulatory orientation and methods, despite the repeated emergence of gendered harms, such as deepfake sexual exploitation and the Iruda chatbot incident. In particular, although the Act establishes safety, trustworthy, and explainability as foundational legal principles, it does not explicitly recognize gender equality or non-discrimination. As a result, gender bias and structural discrimination may not be readily rendered visible as regulatory risks within the legal framework. Moreover, with respect to high-impact AI systems, the statutory criteria for classification remain unclear, and the regulatory scheme—centered primarily on administrative oversight—does not provide adequate mechanisms for remedying the rights of victims of gender discrimination. From a governance perspective, gender equality is addressed only formally in relation to the composition of the National AI Strategy Committee and thus fails to meaningfully incorporate a gender perspective, particularly one grounded in intersectionality, into policy formation and implementation. In light of these limitations, the Framework Act on Artificial Intelligence should be transformed into a regulatory regime that places gender equality at its normative core.