The institute is advancing toward on-site AI action with a robot foundation model, prioritizing safety-centered source algorithms to prevent accidents and enable autonomous, factory-wide intelligent operation.

It envisions a manufacturing AI that autonomously samples data, labels data, and trains models, signaling progress toward autonomous factory operation with the robot foundation model acting as the robots’ brain.

In finance, EXAONE BI is highlighted as an explainable prediction solution expanding globally with partners like LSEG and Koscom, while EXAONE Forecast integrates to handle time-evolving information across broader asset classes, including bonds and commodities.

Exaone BI enables collaborative AI agents for data-driven analysis and, launched earlier this year, is pursuing collaborations with LSEG and Koscom to target global and domestic markets, with Exaone Forecast applied to expand time-series capabilities into new asset classes.

LG AI Research Institute unveiled specialized AI solutions for manufacturing, science, and finance at the LG AI Talk Concert 2026 in Seoul.

At the LG AI Talk Concert 2026, LG AI Research Institute introduced industry-specific expert AIs for manufacturing, finance, and science.

In science, EXAONE Discovery designs materials and predicts synthesis outcomes, with milestones like discovering Ramsidil for hair loss in one day through LG Household & Health Care collaboration, and integrating EXAONE 4.5 with AI-based drug screening with D&D Pharmatech to shorten future drug approvals.

Partnerships include Ramsidil (a hair loss candidate) identified via rapid screening with LG Household & Health Care, and next-generation oral peptide drugs with D&D Pharmatech.

Leaders emphasize solving industrial problems with expert AI that handles many variables and rare exceptions, not just general-purpose models.

A global AI trends keynote and panels highlighted Korean AI insights and ongoing efforts to expand technology to international markets.

Manufacturing foundation models introduced include Exaone Tabular for processing and quality data understanding and Exaone Omni Inspect for robust inspection without retraining despite process or product changes.

Additional manufacturing introductions cover EXAONE Tabular for predicting process conditions and quality, and EXAONE Omni-Inspect for inspection robustness amid changing external images.