With the rapid evolution of artificial intelligence (AI), the concept of AI Drug Discovery has gained significant attention. Headlines often suggest that AI can now “create new drugs” or “automate drug development,” raising expectations in both scientific and business communities.
But how much of this is already realized in actual drug discovery practice? This article offers a practical overview of the current state and limitations of AI Drug Discovery.
Key Drug Discovery Stages Where AI Is Applied
- Target identification
- Hit discovery (compound screening)
- Lead optimization (improving properties and potency)
- Toxicity and safety prediction
- Biomarker discovery (patient stratification)
- Clinical trial design support
Current Levels of Practical Application
AI has shown notable success in hit discovery and lead optimization, particularly in molecular structure prediction and generation by combining machine learning with computational chemistry. Toxicity prediction and ADME profiling also benefit from AI-assisted modeling.
However, AI has not yet revolutionized drug discovery as a whole. It is still far from dramatically improving target selection or clinical trial success rates. In reality, AI functions as a tool supporting specific steps, rather than fully inventing new drugs independently.
How AI Is Used in Practice
For researchers, AI serves as a tool to generate hypotheses and prioritize candidates. Scientists critically evaluate AI-generated suggestions and validate them experimentally. Rather than replacing human expertise, AI provides additional input for informed decision-making.
【My Thoughts】
AI drug discovery holds great promise, but we are still far from having a truly universal “drug invention engine.” AI remains a supportive tool for researchers, where data quality and real-world scientific expertise continue to play crucial roles. Avoiding overhype and applying AI with proper balance will become even more important as the field evolves. Since the landscape is changing rapidly, I will continue to share updated perspectives through this ongoing series.
