[Edaily Reporter Choi Oh-hyun ] LG Corp. is accelerating innovation in industrial settings by leveraging “expert artificial intelligence (AI)” to solve complex challenges in sectors such as manufacturing, finance, and science. Following last year’s announcement of a blueprint to expand into specialized industrial domains—such as agent-based AI and physical AI—LG Corp. unveiled the results achieved by “expert AI” in actual industrial settings this year. LG Corp.’s ultimate goal is to realize an “autonomous factory” where AI understands production processes and robots independently plan and operate the factory.

LG Corp. held the “LG AI Talk Concert 2026” on the 14th at LG Science Park in Magok, Gangseo-gu, Seoul. The event showcased industry-specific expert AI systems and real-world application cases. While last year’s event focused on unveiling the ExaOne ecosystem and its scalability, this year’s event highlighted the tangible results achieved in actual industrial settings.

Lim Woo-hyung, Co-Director of LG Corp. AI Research, emphasized, “What’s more important than AI development and adoption is actually improving work efficiency,” adding, “It must lead to tangible results for companies in industrial settings, such as cost savings and improvements in productivity and quality.”

ExaOne is a multimodal AI model based on a large language model (LLM) that LG Corp. first unveiled in late 2021. LG Corp. is currently expanding its application to specialized fields across various industries. The manufacturing sector is where expert AI is being utilized most actively. LG Corp. employed “ExaOne Tabular,” a model specialized for manufacturing, to analyze structured tabular data generated on the factory floor and predict process status and quality. It also enabled the model to quickly infer changes in process conditions using only a small amount of data. Conventional methods required at least 14 days’ worth of data for model retraining.

In fact, LG Innotek applied Tabular in the early stages of its product manufacturing process, reducing model response lead time by as much as 85%. LG Innotek plans to expand the use of this technology to build autonomous factories for its camera module and printed circuit board businesses. “ExaOne Omni Inspect,” which enables inspection without retraining even when product and process images change, was also introduced as AI specialized for manufacturing sites. LG Corp. plans to apply this technology to actual inspection processes within this year.

By leveraging ‘ExaOne BI’ and ‘ExaOne Forecast,’ the company has expanded its business scope to domestic and international financial markets. ExaOne BI analyzes financial data and provides forecasts and decision-making insights. Forecast analyzes time-series data to predict future values. The London Stock Exchange Group (LSEG) and Koscom are utilizing these tools to analyze stocks listed on the Korean and U.S. stock markets and provide insights to their clients. Arman Sahovic, Head of Data Platform Solutions for the Asia-Pacific region at LSEG, who participated as a panelist in the session held that day, also gave a positive assessment of customer feedback. In fact, the AI-developed ETF, LQAI, is now listed on the New York Stock Exchange and has reached the stage where it is automatically rebalanced every four months.

In the field of science, the “AI Autonomous Laboratory” was unveiled for the first time that day. The Autonomous Laboratory uses AI to predict promising materials, design experiments, and then execute repetitive trials. Robotic arms and laboratory equipment conduct experiments 24 hours a day, and the results are accumulated as data. During this time, researchers can focus on developing new hypotheses and conducting highly complex research.

An example was also presented of how ExaOne Discovery—an AI specialized for science—reduced the time required to discover a new substance from 22 months to just one day. LG AI Research discovered “Ramcidil,” the optimal substance, among 420,000 candidate compounds for preventing hair loss using ExaOne Discovery. LG H&H is currently commercializing products, such as shampoos, that utilize Ramcidil. Using ExaOne Pass, the team analyzed pathological images to design optimal treatment plans for stomach cancer patients. They plan to expand this scope to include lung and colorectal cancers in the future. “ExaOne 4.5” is currently working on a project to be integrated into the Ministry of Food and Drug Safety’s AI system for new drug reviews. Since its establishment in December 2020, LG Corp. AI Research has solved over 100 industrial challenges—including battery life and capacity prediction—and filed 1,080 patents both domestically and internationally.

LG Corp. AI’s ultimate goal is the full automation of factories and research laboratories. When a person sets a goal, the AI understands the factory environment, and robots independently devise and execute work plans. This structure goes beyond the automation of individual equipment to enable the entire factory to operate as a single intelligent entity.

Co-Director of Research Lim Woo-hyung stated, “Over the past six years, LG Corp. has been enabling AI to understand industrial problems, and we are now preparing for the stage where that AI can make its own judgments and take action.” He added, “We will work toward a day when everyone can enjoy the value befitting the name ‘EXaOne’—Expert AI for everyone.”