42Maru (CEO Dong-hwan Kim), a startup specializing in generative artificial intelligence (AI), announced on the 4th that its ultra-large language model 'LLM42' has obtained 'AI-MASTER' certification, a domestic private AI reliability certification system. This is a follow-up achievement to passing the 'AI Reliability Certification (CAT)' of the Telecommunications Technology Association (TTA) for the first time in Korea last year, and 42Maru has become the only company to have the first record of AI reliability certification in both the public and private sectors.
AI-MASTER is Korea's first private AI reliability evaluation system jointly developed by the Korea Artificial Intelligence Industry Association (AIIA), Suresoft Tech, and Selectstar. This certification is based on the EU's seven principles of 'Trustworthy AI' and nine ISO/IEC international standards, and is conducted through document, test, and data-based quantitative evaluations of a total of 63 items. More than 60% of all evaluation items utilize automated tools to enhance objectivity and consistency.
In this evaluation, FortiTumaru's LLM42 received high scores in categories such as privacy protection, bias response, and technical stability. LLM42 is a lightweight, ultra-large language model specialized for Korean, and is characterized by its low cost and high security, which ensures applicability to industrial sites. It improves the hallucination phenomenon, a major issue in generative AI, through search-augmented generation (RAG42) and reading-based generation (MRC42) technologies, and also supports private mode that can protect sensitive data.
Fortitumaru was also the first to receive certification for an ultra-large language model in the 'AI Reliability Certification (CAT)', a public certification system operated by the Ministry of Science and ICT and TTA in November 2023. At that time, it passed precise verification in 11 reliability items, including responsibility, transparency, and safety.
FortiTumaru CEO Kim Dong-hwan said, “In the midst of the spread of generative AI technology, reliability and responsibility are the key criteria for sustainable growth,” and added, “LLM42’s official certification is a case that simultaneously proves the safety and reliability of the technology, and we will continue to focus on providing stable AI infrastructure to the public and industry in general.”
- See more related articles
You must be logged in to post a comment.