en.Wedoany.com Reported - SK Telecom's self-developed AI foundation model will be applied to the manufacturing industry for the first time, in collaboration with steel manufacturer KG Steel and automotive parts manufacturer Konex. On the 25th, SK Telecom announced that it has signed strategic business agreements with both companies to promote the empirical application of AI agents based on its AI foundation model in factory sites.
Starting in April this year, SK Telecom obtained high-quality data from KG Steel and Konex, including past process error and accident analysis reports, equipment manuals, and logs. Based on its self-developed AI foundation model A.X K1, it developed a demonstration version of a "manufacturing-specific AI agent." A.X K1 is a large-scale language model with 519 billion parameters, excelling in handling complex tasks, but activating only approximately 33 billion parameters during inference. This structure, where the overall model is large but only necessary parts are activated, enables efficient operation in industrial settings.
SK Telecom, KG Steel, and Konex plan to apply this demonstration version for field verification in the second half of this year at KG Steel's Dangjin plant's cold rolling line and Konex's casting and machining processes. During the verification, KG Steel and Konex will share more high-quality manufacturing process data with SK Telecom in real time; SK Telecom will improve the AI agent's performance, inference speed, and expand its functions based on this data and field feedback. Additionally, manufacturing site data obtained during the verification will be used to train the A.X K2 model currently under development. After the verification is completed, the three parties will consider commercializing and introducing the "manufacturing-specific AI agent," and if necessary, study the possibility of replacing the model with subsequent series of SK Telecom's self-developed AI foundation model.
The manufacturing industry has long been considered a field where AI is difficult to introduce. The digitization of manufacturing site data has been slow, and data is often generated and managed separately by process or department, making it difficult for AI to utilize. Additionally, the work methods of operators vary based on skill level and experience, and the "knowledge isolation" phenomenon, where core know-how exists only with specific skilled workers, is a typical challenge. SK Telecom stated that by digitizing and assetizing scattered data and skilled workers' experiential knowledge, and introducing AI agents that learn this knowledge to the site, it is expected to quickly respond to process errors, shorten processing time, and improve efficiency.
Jung Seok-geun, head of SK Telecom's AI CIC, stated that for manufacturing sites where safety is critical, a self-developed AI foundation model that can be utilized without data leakage is an effective solution. He added that the company will accelerate AI transformation in manufacturing and expand application cases of its self-developed AI foundation model, starting with the collaboration with KG Steel and Konex. Bae Seon-woo, head of KG Steel's Technology Research Institute, emphasized that this collaboration lays the foundation for AI introduction based on site data, which will further enhance manufacturing competitiveness. Lee Gwang-pyo, CEO of Konex, pointed out that quickly responding to recurring quality issues on site is a long-term challenge for manufacturing, and AI will help improve work efficiency at manufacturing sites.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
