Part 5/9:
The process used by MIT involves generative models. These models explore vast chemical spaces to identify promising molecular structures. In practice, this means:
Creating endless variations of compounds
Predicting which will work effectively
Validating these predictions through lab tests
This approach drastically cuts down traditional drug development timelines, which historically span 10 to 15 years and cost upwards of $2 billion per drug. With AI, these timelines could be shortened to less than 5 years, even around 2-3 years for certain drugs.
Step-by-Step Transformation of the R&D Pipeline
AI’s impact isn't limited to molecule design. It extends to every phase of the drug lifecycle: