Samsung Electronics has begun full-fledged development of its own Large Language Model (LLM) for internal use. The South Korean tech giant has invested a lot of resources and manpower in this development, which is being led by Samsung Research. The company aims to complete the development of the first version of its LLM within two months.
Samsung plans to use the LLM for document summarization, software development, and language translation. However, the South Korean firm has not yet decided whether to make its AI solution available for general consumers. The company initially relied on third-party generative AI solutions, but its internal data was leaked when using them. As a result, the company has started developing its own AI solution.
Samsung believes that such AI solutions can significantly reduce the time required for software development and semiconductor design. The company was engaged in talks with firms such as Google, Microsoft, Naver, and OpenAI, but its top-level management meeting, attended by Lee Jae-yong, decided that Samsung should develop its own AI solution.
Currently, various South Korean firms are developing their own LLM-based generative AI solutions, including Kakao’s KoGPT, LG‘s ExaOne, Naver’s HyperCloverX, and SKT’s A-Dot. These solutions can draw and write like humans, but the accuracy of data and information privacy remains a hurdle.
Training LLMs requires powerful graphics processing units (GPUs), and Samsung is using its entire GPU resources for LLM development. The company aims to use the LLM for faster document summarization and software development, but it remains to be seen whether the resulting AI solution will be made available to the public.
