Child Protection in the Digital Age in South Korea: Professionals' Perceptions on Artificial Intelligence (AI) and Big Data Efficacy and Ethical Challenges

Background & Purposes: Recent advancement in artificial intelligence (AI) and big data analytics have transformed child abuse detection by enabling the analysis of various data sources, offering benefits like early risk identification and enhanced decision-making. However, ethical and practical considerations regarding data privacy, algorithmic bias are paramount for the responsible and effective implementation. This study investigates the current utilization of AI and big data for child abuse prediction and prevention in the child welfare system of South Korea, assessing their effectiveness through the lens of professionals and exploring associated ethical implications. This study fills a gap in existing literature by examining diverse perspectives of child protection professionals, offering valuable insights for technology-driven efforts to safeguard vulnerable populations.

Methods: Data were collected through semi-structured, in-depth qualitative interviews with professionals with expertise in child protection in South Korea, particularly those engaged in AI and big data initiatives. A combination of purposive recruitment and snowball sampling techniques was utilized to ensure the inclusion of professionals best suited for the study. Ten participants were involved in the study, and thematic analyses were employed to identify emerging patterns and themes from the data. The analysis followed a structured approach to maintain consistency and accuracy in theme identification. Each researcher independently reviewed all transcripts, identifying themes. Subsequently, the researchers reconvened to collectively discuss and further refine the identified themes. All interviews were audio-recorded and transcribed verbatim.

Results: The themes are extracted to understand professionals’ perceptions regarding the utilization and limitations of AI and big data in child protection. Participants emphasized the importance of foundational data and the potential of AI algorithms in uncovering cases of re-abuse. However, concerns were raised about the reliability of predictive variables, bias in case selection, and the need for qualitative data to enhance prediction accuracy. While participants acknowledged the system’s predictive power, they highlighted its low accuracy and lack of integration with existing services as limitations. Challenges were reported regarding the lack of enforceability of policies and a shortage of child protection workers. Furthermore, ethical considerations, such as privacy issues and the societal culture of blaming, were discussed as significant challenges.

Conclusion & Implications: The study underscores the complex interplay among technological advancements, ethical considerations, and societal dynamics in utilizing AI for child welfare. While AI algorithms demonstrate potential in identifying crisis situations, their effectiveness is contingent upon addressing issues of bias, integrating qualitative data, and ensuring ethical data utilization. Furthermore, efforts to raise awareness, strengthen education, and bridge gaps between policy and practice are essential for maximizing the benefits of AI while mitigating its limitations in safeguarding children and families at risk.

While this study centers on South Korea, its findings carry implications for child welfare practices globally. Through its examination of AI and big data for child abuse prediction and prevention, the study sheds light on strategies and challenges relevant across diverse cultural contexts, potentially enhancing child protection efforts worldwide.