Comparative Analysis of Generative AI Risks in the Public Sector

The landscape of artificial intelligence (AI) has experienced a monumental shift with the emerging of Generative AI (GenAI), which has demonstrated to be a transformative tool across diverse sectors. GenAI outputs can span various digital formats, including text, images, videos, and audio, generating particular interest in the public sector. The growing interest of governments in integrating GenAI technologies in public sector operations is marked by the creation of emerging governance instruments and the formulation of soft laws, like standards, principles, and guidelines. This study aims to delve into the intricacies and potential risks associated with the deployment of GenAI within government. Through a qualitative content analysis, the research meticulously examines GenAI usage guidelines issued by Australia, Canada, New Zealand, the United Kingdom, and South Korea. The objective is to discern the risks acknowledged by these countries' soft laws and compare them with the risks identified by scholars in the field. The performed comparative analysis across countries suggest that the use of GenAI in the public sector raises common risks such as information leakage, data privacy, security, and concerns over public trust. By elucidating the varied risk perceptions across different national contexts, this study provides theoretical and practical implications related to the risks of GenAI within the public sector. Moreover, it sets a foundation for future research and policy development, ensuring that generative AI is used as a force for good in public governance.