Artificial Intelligence (AI) is transforming the search for solutions to complex diseases. This is the fourth article in a six-part series exploring how AI is reshaping medical research and treatment methods.
During a video call, Alex Zhavoronkov, CEO of Insilico Medicine, holds up a small green, diamond-shaped pill. His company developed the pill to treat idiopathic pulmonary fibrosis (IPF), a progressive lung disease with no known cause or cure. While still awaiting approval, early clinical trials have shown promising results.
This drug represents a new wave of treatments where AI has been critical to the discovery process. “We may not yet have the first AI-designed molecule approved,” says Dr. Zhavoronkov, “but we are among the closest.” Insilico Medicine is one of many players in the growing field of AI-driven drug discovery.
A New Era in Drug Discovery
AI is revolutionizing the pharmaceutical industry, traditionally dominated by medicinal chemists. Both specialist AI-driven biotech firms and established pharmaceutical companies are embracing AI. Even tech giants like Alphabet, Google’s parent company, have entered the space. Alphabet launched Isomorphic Labs, an AI drug discovery firm, in 2021, led by Nobel Prize winner Demis Hassabis.
Using AI could drastically reduce the time and cost of drug development. Currently, bringing a drug to market takes 10–15 years and costs over $2 billion. Additionally, nearly 90% of drugs fail during clinical trials. By employing AI, researchers hope to identify potential drugs more efficiently and increase the success rate.
Professor Charlotte Deane of Oxford University believes AI’s integration into drug discovery marks the start of a transformative era. However, human collaboration remains essential. AI tools enhance the work of scientists rather than replace them. A study by Boston Consulting Group (BCG) shows that over 75 AI-discovered molecules are already in clinical trials, with many more in development.
How AI Streamlines Drug Development
AI’s impact is most evident in two key areas: identifying therapeutic targets and designing molecules. Traditionally, scientists identify disease-related genes or proteins experimentally. AI, however, analyzes massive datasets to uncover molecular links and suggest targets more efficiently.
In molecule design, generative AI—similar to the technology behind ChatGPT—creates potential drug candidates. This approach replaces the labor-intensive process of synthesizing and testing hundreds of molecules. For instance, Insilico Medicine used AI to design a molecule targeting TNIK, a protein linked to IPF. The process took just 18 months and involved testing 79 molecules, compared to the industry average of four years and 500 molecules.
Despite these advancements, challenges remain. Limited data can hinder AI training and introduce biases. Companies like Recursion Pharmaceuticals address this by conducting automated experiments to generate extensive datasets. Recursion developed a molecule for lymphoma and solid tumors using AI to find a new way to target a specific cancer-driving gene. This molecule is now undergoing early clinical trials.
The Road Ahead
While AI shows immense potential, its ultimate success depends on proving that AI-discovered drugs can pass clinical trials and outperform traditional methods. Chris Gibson, CEO of Recursion, emphasizes the need for consistent results. “When AI-discovered drugs demonstrate higher success rates,” he says, “the world will fully embrace this approach.”
AI is already reshaping drug discovery, cutting costs, saving time, and inspiring hope for better treatments. As the technology evolves, its role in transforming medicine will only grow.
