"Global AI drug discovery accelerates…KOR faces talent·data shortage"

Global big tech + big pharma partnerships are advancing AI-driven drug discovery

Dr. Pyo Junhee, "Fully automated era of data-experiments, beyond particular research step"

The South Korean pharmaceutical industry is facing hurdles of talent and data shortage

The global drug discovery ecosystem is restructuring centered on artificial intelligence (AI). Partnerships between big tech and big pharma companies are driving rapid advancements across the entire drug development lifecycle.  

However, the Korean biopharmaceutical industry continues to face significant hurdles, including a shortage of skilled talent and limited access to data.

Dr. Pyo Junhee, Director of the AI Institute for Drug Development

On the afternoon of the 2nd, Director Pyo outlined the accelerating AI drug discovery ecosystem during a workshop titled "Future trends of AI technology in the biopharmaceutical industry," which was jointly hosted by the Pharmaceutical Society of Korea (PSK), the Korean Society of Pharmaceutical Sciences and Technology (KSPST), and the Korean Academy of FDC Regulatory Science.

During her presentation, Pyo emphasized that among the pipelines currently under development by global big pharma companies, it has become difficult to find a novel drug candidate that does not integrate AI.

Big pharma companies are staking its future on strengthening internal capabilities, aggressively recruiting AI engineers and data scientists for 100 to 200 positions.

The growth trajectory is further accelerated by the interest and capital injection from big tech companies that view AI-driven drug discovery as a primary growth engine. Consequently, there has been a massive influx of "bio-sequence Large Language Models (LLMs)" capable of decoding and understanding biological sequences such as DNA and RNA.  
"Nvidia announced a partnership with Eli Lilly to build an AI drug discovery factory, while Google DeepMind developed AlphaFold3 and recently unveiled enhanced iterations of the model," Pyo explained. She noted that when Nvidia initially launched its 'BioNeMo' platform, several Big Pharma companies reportedly paused physical laboratory experiments to run toxicity predictions through BioNeMo, subsequently recalibrating their pipeline priorities.  

Furthermore, Anthropic recently introduced 'Claude for Science,' which offers comprehensive end-to-end workflows tailored for drug discovery. Using a multi-agent AI framework, a coordinator agent categorizes a research task, after which specialized sub-agents, focusing on genomics or structural biology, independently validate and integrate their findings.

Pyo emphasized that the industry has shifted beyond merely accelerating isolated research steps. Realizing an era where hypothesis generation, experimental design, and data analysis are fully automated has already become a reality.

Accordingly, Pyo stressed that the industry is no longer competing on the speed of localized workflows, but has officially entered an era of intense competition over de novo structural design.

However, the Korean pharmaceutical industry continues to face bottlenecks, most notably a severe shortage of skilled personnel, as well as data scarcity and data quality issues.

​"We conducted a survey to identify the specific pain points of domestic pharmaceutical companies and AI drug discovery firms," Pyo stated. "The most prevalent response was a shortage of proficient talent and recruitment difficulties, followed closely by data availability and quality concerns."

Because proprietary drug discovery data represents a highly valuable corporate asset for individual companies, it is rarely generalized and remains highly fragmented or siloed across the industry.

To address this, Pyo suggested that "data partnerships structured around commercial transactions or mutual research needs may be selectively feasible when financial incentives align. Additionally, we are seeing the emergence of specialized consortia aimed at solving common clinical goals, such as sharing control-arm data from clinical trials."

The South Korean government is also actively investing in AI-driven pharmaceutical advancement through initiatives such as the K-AI Drug Discovery Preclinical and Clinical Model Development Project.  

In late 2025, the Ministry of Health and Welfare (MOHW) allocated approximately KRW 37.1 billion to the 'K-AI Drug Discovery Preclinical and Clinical Model Development Project', appointing lead operating organizations to foster a AI-driven development ecosystem. The Korea Pharmaceutical and Bio-Pharma Manufacturers Association (KPBMA), which houses the AIDD, was designated as a primary lead institution.  

​"Currently, three major hospitals, pharmaceutical corporations, and research institutions are collaborating to build integrated datasets that bridge the preclinical and clinical phases," Pyo said. "Using these datasets, we are developing foundation models and a diverse array of downstream AI applications."  

Pyo concluded by stating, "For clinical complexities that AI alone cannot fully resolve, we are integrating various predictive simulation models to develop a platform capable of optimizing clinical trial designs."