LG's EXAONE Discovery AI model to be applied to chemical research

Shin Seok-min (left), president of the Korea Research Institute of Chemical Technology, and Lim Woo-hyung, co-director of LG AI Research, hold up a signed agreement at an MOU ceremony between the two institutions. [Korea Research Institute of Chemical Technology]
Shin Seok-min (left), president of the Korea Research Institute of Chemical Technology, and Lim Woo-hyung, co-director of LG AI Research, hold up a signed agreement at an MOU ceremony between the two institutions. [Korea Research Institute of Chemical Technology]

A research and development model in which AI identifies candidate chemical compounds and scientists verify them through experiments is set to move forward in earnest.

The Korea Research Institute of Chemical Technology said Wednesday it signed an MOU with LG AI Research to strengthen cooperation between chemistry and AI.

The agreement aims to combine the institute's accumulated expertise and experimental infrastructure in chemistry with LG AI Research's AI technology to accelerate the development of new materials and improve research efficiency.

The two institutions will collaborate in three areas: joint AI research in chemical R&D, standardization of chemical research data, and identification of new research projects that integrate chemistry and AI.

Central to the partnership is applying EXAONE Discovery — a chemistry-specialized AI model developed by LG AI Research — directly to the institute's research operations.

EXAONE Discovery uses AI to search for new chemical compound candidates and predict material properties. Rather than having researchers synthesize and test large numbers of candidates one by one, the model identifies the most promising options first.

The institute plans to verify AI-suggested candidates through actual experiments and use the resulting research findings to improve the AI model.

The collaboration is designed to cut the time and cost required to develop new materials while improving the efficiency with which difficult problems in chemistry are solved.

Participants pose for a photo at the MOU signing ceremony between the Korea Research Institute of Chemical Technology and LG AI Research. [Korea Research Institute of Chemical Technology]
Participants pose for a photo at the MOU signing ceremony between the Korea Research Institute of Chemical Technology and LG AI Research. [Korea Research Institute of Chemical Technology]

The two institutions also plan to develop an AI research assistant agent to support scientists in the field. Alongside upgrading the AI model itself, they intend to help researchers use AI when analyzing chemical data and exploring new research directions.

Standardizing chemical research data is another key area of cooperation.

The two institutions will work on AI-based approaches to data standardization in preparation for the Act on the Management and Promotion of National Research Data, which is set to take effect in June 2027.

The institute plans to strengthen its research data management and utilization capabilities with advisory support from LG AI Research. To that end, the two sides will form a working-level consultative body and conduct regular reviews of progress on joint projects.

"The competitiveness of chemical research depends on how well accumulated research data and experimental verification capabilities are combined with AI," said Shin Seok-min, president of the institute. "We will work with LG AI Research to accelerate solutions to difficult challenges in chemistry and prepare for a new leap forward in realizing KRICT Vision 2040."

"A candidate compound proposed by AI only becomes truly valuable in the research field when it is verified through experiment," said Lim Woo-hyung, co-director of LG AI Research. Lim added that combining EXAONE Discovery's predictive capabilities with the institute's experimental and verification infrastructure would raise both the speed and success rate of new material discovery.


nbgkoo@heraldcorp.com