Generative AI startup 42Maru (CEO Dong-Hwan Kim) announced on the 30th that it had introduced its AI application strategy in the medical and healthcare industry during a keynote speech at the '2026 Joint Symposium' held at Sonocam Vivaldi in Chuncheon, Gangwon-do for two days from the 29th. This symposium was co-hosted by the Department of Big Data Medical Convergence and the Department of Data Science at Kangwon National University, and was designed to discuss key issues in the field such as medical AI security and explainable AI (XAI).
In the keynote session on the first day of the event, CEO Dong-Hwan Kim explained how agentic AI, which makes decisions and performs actions on its own, can improve the workflow in the medical field under the theme of "The Age of Agentic AI – Care Innovation in the Era of Generative AI." He introduced global AI healthcare trends and emphasized that AI is being applied to practical medical service areas such as diagnosis, prediction, medical record automation, and patient-tailored care through overseas cases such as Google's medical-specialized LLM Med-PaLM and MedGemma, and Microsoft's AI diagnosis orchestrator MAI-DxO.
CEO Kim shared domestic application cases such as ▲AI-based hospital work efficiency improvement ▲automatic creation of clinical treatment charts based on patient data ▲AI-based nationwide psychological care counselor assistance service ▲personalized emotional support and care service as part of Forti2Maru's strategy, and presented a step-by-step roadmap for building a global-level healthcare specialized model and internalizing clinical work.
He particularly emphasized the need to develop a foundation model specifically tailored to the Korean medical and healthcare sector. He explained that the medical field faces functional limitations with a general-purpose foundation model alone, requiring a comprehensive approach from architecture design, FromScratch development, and the operation of a sovereign AI-based model that takes into account the domestic regulatory environment and data specificities. He also emphasized the importance of building a model that reflects the physical and emotional characteristics of Koreans.
FortiTumaru mitigates the hallucination problem of ultra-large language models through RAG42 and MRC42 technologies. It also develops a specialized, lightweight model, LLM42, for industry use, providing a private mode for enterprises. This not only protects internal corporate data and sensitive information, but also reduces the cost of building AI solutions.
CEO Kim Dong-hwan said, “As medical and healthcare involve a lot of sensitive data, strong security and trustworthy design are essential,” and added, “Forty2Maru will implement AI care services optimized for domestic medical settings with lightweight models specialized for healthcare and RAG engineering technology.”
- See more related articles
You must be logged in to post a comment.