According to a report, Samsung Electronics has begun full-scale development of its own Wide Language Model (LLM) for internal applications after its internal data was recently leaked. Led by Samsung Research, the project received significant investment in manpower and resources. In particular, all available GPU resources are devoted to the development of LLM, which is an important component in training such models. Samsung aims to complete the first version of the LLM within two months.

Samsung’s AI solution aims to reduce time needed to develop software

Samsung’s LLM solution will focus on document abstraction, software development and language translation. While the company has not yet decided to make its AI solution generally available, its internal use is seen as a priority. Previous reliance on third-party productive AI solutions, especially ChatGPT, led to an unfortunate data leak, prompting Samsung to ban such third-party solutions and start building its own AI technology.

Samsung believes that deploying AI solutions based on LLM technology can significantly reduce the time required for software development and semiconductor design. As we explored partnerships with various well-known companies such as Samsung, Google, Microsoft, Naver and OpenAI, the final decision to develop an in-house artificial intelligence solution was taken at a senior management meeting that included Lee Jae-yong.

A number of South Korean companies, including Kakao, LG, Naver and SKT, are also developing their own LLM-based productive AI solutions, such as KoGPT, ExaOne, HyperCloverX and A-Dot. These solutions exhibit human-like drawing and writing abilities. However, issues with data accuracy and information privacy remain serious obstacles that need to be addressed in the current development process.