LG Uplus and LG AI Research Institute held a meeting on the 5th (local time) in Barcelona, Spain, where Mobile World Congress (MWC) 2025 is taking place, and introduced their jointly promoted 'One Team AI' strategy. Attendees included Lee Sang-yeob, Chief Technology Officer (CTO) of LG Uplus, and Lee Hong-rak, Chief AI Scientist and Vice President of LG AI Research Institute.
The two companies explained the main technologies and collaboration cases that constitute human-centered AI, the first step in achieving the goal of 'creating a bright world through human-centered AI,' and announced their long-term AI advancement plan.
Lee Sang-yeob, CTO of LG Uplus, noted that 'LG Uplus and LG AI Research Institute will present AI services that provide 'reliable' and 'customized convenience' to customers through close collaboration, functioning as a single team,' and said they would lead to a bright world powered by human-centered AI by providing better experiences for customers through AI technology.
LG Uplus and LG AI Research Institute are advancing 'On-device AI' technology that can completely prevent customer data leakage in order to create AI that 'helps people, provides convenience, and at the same time maintains safety and trust.' The 'ixi-O' service, which applies the On-device AI technology where customer data is not stored on servers, is the first in the domestic mobile telecommunications industry.
The On-device AI technology is currently implemented in the 'ixi-O' service. LG Uplus and LG AI Research Institute are developing 'On-device sLM' that uses Neural Processing Units (NPU) by further enhancing existing On-device AI technology that utilizes Central Processing Units (CPU). The ixi-O service, which applies the NPU-based small language model (sLM), can reduce power consumption to one-fourth and CPU usage to one-tenth compared to the current levels, thereby improving efficiency and performance.
'A high-performing sLM' is essential for On-device sLM. LG Uplus is conducting optimization work specialized for telecommunications services based on EXAONE, developed by LG AI Research Institute, which has been recognized globally for its outstanding performance among lightweight models. LG AI Research Institute plans to develop a model that, while reducing the model size to half compared to the latest version, EXAONE 3.5, will provide even better performance to be embedded in the ixi-GEN.
With the implementation of the On-device sLM, the ixi-O service is expected to offer more personalized services through enhanced computational capabilities. As a result, the ixi-O is anticipated to evolve into an 'Actionable AI' service capable of not only call recording and summarization but also scheduling appointments and making restaurant reservations.
To achieve 'human-centered AI,' LG Uplus and LG AI Research Institute are also focusing on 'AI ethics.' As the pace of AI development accelerates, AI ethics, which has emerged as a key issue, plays a role in providing guidelines to ensure trustworthiness and fairness, as well as preventing the misuse of AI and ensuring that technology is used positively.
LG Uplus has established an organizational operation and compliance system at the LG Group level to fulfill its ethical responsibilities regarding AI technology in collaboration with LG AI Research Institute. In particular, LG AI Research Institute has built an AI ethics governance framework to supervise and manage the entire process of AI research and development and usage, ensuring that incorrect decisions are not made, while serving as a hub that establishes AI ethics-related policies across all subsidiaries of LG Group.
To create the most ethical AI, LG AI Research Institute has also developed a 'Data Compliance Agent' that automatically analyzes the legal risks of data using AI. Previously, human verification was necessary for the risks associated with the data used for AI learning, but by utilizing technology from LG AI Research Institute, it is possible to make more accurate and cost-efficient judgments about the risks of AI training data. This enables LG Group to automatically analyze and evaluate data that could lead to significant opacity and legal risks, thus facilitating the development of more trustworthy AI services.
LG Uplus is also strictly adhering to AI ethics in accordance with the ethics governance of LG AI Research Institute, while implementing its own measures to enhance AI ethics. LG Uplus is working to ensure that AI services are provided ethically by anonymizing input information even if customers unconsciously enter personal data, automatically blocking potentially harmful information, and refusing to respond to violent or inappropriate questions.
Lee Hong-rak, Chief AI Scientist and Vice President of LG AI Research Institute, said, 'LG AI Research Institute will secure a global level of technological capability from cutting-edge research to model development and will create a human-centered agent AI through this.' He added that through the 'reliable and trustworthy' ixi-O, LG Uplus and the team will expand synergies so that customers can utilize differentiated, customized AI services.