South Korean tech firms develop specialized AI to ensure data sovereignty

South Korean technology firms are intensifying efforts to secure AI sovereignty through specialized large language models (LLMs) tailored to domestic regulatory and security needs.

Bit Computer has partnered with Amazon Web Services (AWS) to develop a customized medical LLM. This initiative aims to address the limitations of global AI models in handling South Korea’s unique medical billing systems, complex prescription codes, and strict data privacy regulations. By training models on Bit Computer’s 43 years of de-identified medical data within the AWS cloud, the collaboration seeks to automate clinical documentation and administrative tasks while ensuring sensitive patient data remains secure.

Simultaneously, Naver Cloud is leading a major consortium to develop a specialized cybersecurity AI foundation model. Supported by government initiatives, Naver Cloud plans to invest significant computing resources, including 4,000 GPUs, to build a 300B-parameter model designed to detect and block cyber threats. Working alongside partners like LG AI Research, the project utilizes a dual-track approach where attack models identify vulnerabilities and defense models learn to mitigate them, aiming to reduce reliance on foreign AI technologies for national security.