How I use AI to turn failed drugs into new medicines

Layla Hosseini-Gerami is using artificial intelligence to repurpose failed drug candidates into new medicines, as detailed in a Nature publication on June 10, 2026. Her approach integrates computational chemistry and biological data with AI modeling to identify existing compounds that could be effective against different diseases. This method aims to accelerate drug discovery by leveraging the extensive research already invested in compounds that did not succeed in their original trials. Hosseini-Gerami's work focuses on identifying therapeutics with "huge potential" that were previously overlooked or abandoned due to efficacy or safety issues in their initial development phases. By applying advanced AI algorithms, her team can analyze vast datasets to predict new therapeutic applications for these legacy drugs. This innovative strategy could significantly reduce the time and cost associated with bringing new treatments to market, as the foundational research and safety profiles of these drugs are already established. The process involves sophisticated AI models that can predict molecular interactions and biological pathways, enabling the identification of novel therapeutic targets for repurposed compounds. This research highlights a promising avenue for pharmaceutical innovation, addressing the high attrition rates in traditional drug development pipelines. The goal is to unlock the value of previously unsuccessful drug candidates, potentially leading to breakthroughs in treating various conditions.