Search for a command to run...
With the rapid advancement of artificial intelligence (AI) technologies and the diversification of commercial models, the demand for AI adoption in the public sector has surged dramatically. Governments and public institutions are actively leveraging AI technologies to enhance administrative efficiency, deliver personalized services, and enable data-driven policy-making. However, in the actual implementation process, numerous limitations and challenges have emerged. In particular, when AI systems are designed in a way that is dependent on specific cloud platforms or proprietary models, organizations are often burdened with high costs and technical constraints during system transitions, expansions, and maintenance in response to evolving technologies. In such a fast-changing AI ecosystem, ensuring the stability and scalability of AI services in the public sector requires establishing cross-compatibility across AI infrastructure, cloud platforms, and AI models. Cross-compatibility refers to the ability for services, data, and models to be seamlessly interoperable and reusable across different AI environments. This enables public institutions to avoid vendor lock-in and adopt a diverse range of suppliers and solutions. Moreover, securing cross-compatibility acts as a strategic means to reduce long-term maintenance costs while strengthening the sustainability and technological autonomy of public AI services. Nevertheless, many current AI projects in the Korean public sector are being implemented based on heterogeneous standards, model formats, and platform environments. As a result, compatibility issues frequently arise during model replacement or technology migration processes. To effectively introduce and manage diverse AI technologies, it is imperative for public institutions to adopt cross-compatibility strategies grounded in platform independence and open standards. Therefore, this paper proposes design-level requirements aimed at securing cross-compatibility in AI services for the public sector.