AI Discovers New Drugs, Robots Conduct Experiments... Building a Self-Evolving Laboratory

Ministry of Science and ICT Establishes AI Drug Development Hub in Daegu
Utilizing Data from Two Million Patients

When artificial intelligence (AI) identifies new drug candidate compounds, an automated laboratory verifies them, and the results are further utilized to train the AI, establishing a "cyclical new drug development" system in Korea. By integrating data from around 2 million patients with AI and automated laboratories, the goal is to reduce both the time and cost required for discovering and validating new drug candidates.

On August 12, the Ministry of Science and ICT announced the launch of the "AI×Bio Innovation Research Hub Demonstration Project Group in Synthetic New Drug Sector" in Daegu and initiated efforts to establish an AI-based new drug development innovation hub.

Photo material to aid article understanding. A researcher at a pharmaceutical company is searching for new drug candidate compounds through an AI-based new drug development system. The Asia Business Daily DB

This project represents the government's first innovation research hub in the "K-Moonshot" AI new drug development initiative. Key participants include Kyungpook National University, Kyungpook National University Hospital, the Daegu-Gyeongbuk Advanced Medical Industry Promotion Foundation, Uniba, and other universities, hospitals, research institutions, and companies.

The centerpiece is the "Lab-in-the-Loop" system, which connects AI models with real laboratories. AI analyzes bio-data to propose new drug candidate compounds and experimental conditions, which are then validated by the automated experimental platform. The resulting experimental data are fed back into the AI, further improving its prediction capabilities. This process is repeated: "AI prediction → automated experiment → data generation → AI retraining."

Although research utilizing AI for discovering new drug candidates has been ongoing, this project is distinct in that it establishes a closed-loop research system where AI predictions are directly linked to real-world experiments and the results are subsequently used to retrain the AI. The aim is to create a persistent research environment in which AI and laboratories interact continuously, evolving beyond AI as a mere support tool for researchers.

Connecting Patient Data from 2 Million People and Automated Laboratories

Kyungpook National University is responsible for overseeing the project and operating the AI-based synthetic drug Lab-in-the-Loop research system. The university will also establish computing infrastructure for developing AI models and foster talent specializing in AI-bio convergence.

Kyungpook National University Hospital will leverage patient-based data on around 2 million individuals to create a system for utilizing medical data essential for AI-driven drug development.

The Daegu-Gyeongbuk Advanced Medical Industry Promotion Foundation, in collaboration with HLB Life Science, AronTier, Molcube, and Able Labs, will establish high-speed validation infrastructure. Plans are in place to verify AI-proposed candidate compounds and research hypotheses within "A2"-level automated laboratories. The A2 stage refers to semi-automated processes in which multiple sequential steps of synthesis and property analysis are automated, while system setup, maintenance, and some experiments still require human intervention.

Uniba is tasked with developing foundation models for application in drug development, as well as creating a research environment in which newly generated data from experiments are continuously incorporated into AI models.

The Ministry of Science and ICT plans to first validate the Lab-in-the-Loop system in the synthetic new drug sector, where AI technology is well suited for application, and gradually expand it to other AI-bio domains. From next year, the government aims to build additional hubs based on the AI models, data, experimental infrastructure, and operational experience accumulated at the demonstration site.

Oh Daehyun, Director General for Future Strategic Technology Policy at the Ministry of Science and ICT, stated, "The future competitiveness of new drug development will depend on how effectively AI technology and data can be connected to, and utilized in, real experimental environments. Next year, based on the results from the demonstration hub, we will expand into additional AI-bio areas to help Korea lead global AI-bio innovation."

This content was produced with the assistance of AI translation services.