The administration of Lee Jae-myung, which professes to be an 'Artificial Intelligence (AI) government', will actively support AI-based drug development, which is currently receiving substantial investments in the global industry. AI has emerged as a key tool throughout the drug development process, from discovering therapeutic candidates to clinical trials and personalized treatments. The government plans to grow this sector in order to reduce the drug development time from over 10 years and the expense of approximately 2 trillion won to less than half.
The global pharmaceutical and biotech industries are focusing on AI because it can learn vast amounts of data, including compound information and disease-related protein structures, to identify candidates much faster and more accurately than humans. According to the Korea Health Industry Development Institute (KHIDI), the global market for AI drug development is expected to grow from $927 million (1.2651 trillion won) in 2023 to $2.8936 billion (4.553 trillion won) by 2028.
◇Google and NVIDIA enter AI drug development
On the 4th, according to the political circle, the Health and Welfare Committee of the National Assembly recently approved a supplementary budget related to AI drug development prepared jointly by the Ministry of Health and Welfare and the Ministry of Science and ICT. This budget includes projects for 'AI preclinical and clinical model development' and 'development and verification of antibody biobetters utilizing AI models.'
First, 2.184 billion won has been newly allocated for the K-AI preclinical model development project. The project aims to develop a platform and technology that supports the process from the discovery of candidates to clinical trials using AI. The government plans to invest a total of 49.5 billion won from this year until 2029.
Another new project, the development and verification of antibody biobetters, has been allocated 3.3 billion won. Biobetters are new drugs that improve the efficacy or administration convenience of existing biologics, and the core tasks involve discovering candidates and conducting clinical trials using bispecific antibody technology. This project will be promoted at a scale of 40.4 billion won by 2027.
The background behind the government's support policy is the recent global investment surge in the AI drug development sector, amounting to trillion won. In particular, global IT giants such as Google, Amazon, and NVIDIA are making significant investments in AI drug development, leading to rapid growth in the industry.
Alphabet, the parent company of Google, has a drug development corporation called Isomorphic Labs, which entered into a joint development contract with the American pharmaceutical company Eli Lilly worth 2.4 trillion won and with the Swiss company Novartis worth 1.7 trillion won last year. The American AI drug development corporation Recursion Pharmaceuticals acquired the British company Exscientia for about 1 trillion won.
◇Domestic AI drug development still in early stages
Conversely, the domestic AI drug development industry is still in its early stages. Currently, it is estimated that only about 10 drug candidates have entered clinical trials through AI platforms. Most remain in phase 1, with the only one completing phase 2 being Innovo Therapeutics, which is developing the topical scar treatment 'INV-001'.
Moreover, most are currently recording deficits. Innovo Therapeutics, which has the fastest development speed, incurred a deficit of about 13 billion won last year, while Pharos iBio, which was the first domestic AI drug development company to enter clinical trials, reported a loss of approximately 12.6 billion won last year. Standigm, which garnered attention after receiving its first overseas investment, recently reduced its workforce from 80 to 27 through restructuring.
There have been no cases where candidates identified by AI have actually been commercialized. Contrary to expectations, many failures have emerged because AI is primarily utilized only in the preclinical stages of drug development.
To bring a new drug to market, candidate substances must first be identified, followed by preclinical studies to verify safety. After that, clinical trials involving human subjects must demonstrate efficacy and obtain approval from health authorities. This process typically takes around 10 to 15 years.
While AI is said to be able to reduce this timeline by half, the technology for AI drug development currently only assists for a few years during the initial candidate discovery phase. An official from an AI drug development company noted, "While AI can play a significant role in quickly identifying candidate substances, the actual clinical trials and approval process still present significant bottlenecks that take a great deal of time," adding, "There is a need for technology advancement that enables AI to not only play a role in the discovery phase but also in subsequent clinical trials."
In particular, using AI for drug development requires providing data, but there are fundamental limitations. In drug development, actual patient data becomes a key resource after phase 2 of clinical trials. However, access to clinical trial data from hospitals in Korea is low, and the number of patient clinical trial figures corresponds to sensitive personal information, making it difficult to utilize.
Kim Yong-woo, head of the Korea Health Industry Development Institute (KHIDI), stated, "In order for corporations to demonstrate their technological competence, an environment needs to be established where hospital data can be utilized more extensively, but the current situation in the country is still lacking." He expressed hope that the government's support for AI drug development would expand beyond initial candidate discovery to full-scale clinical trials and commercialization, leading to enhanced data accessibility for corporations and strengthened research and development capabilities.