The use of artificial intelligence (AI) in pharmaceutical research allows for a significant acceleration in the search for and creation of new molecules. This was stated by Petr Rodionov, general director of Geropharm, at the Bioprom forum.
According to him, employing AI models in the early stages of research helps substantially reduce the number of compounds that need to be synthesized and tested in the laboratory. “In the development of an innovative insulin molecule, we generated around a million new compounds. An AI model helped us select a few dozen, which we then used for further preclinical studies. Previously, without AI, it was necessary to synthesize thousands of compounds, but now we’re talking about dozens,” Rodionov explained.
He emphasized that the use of AI and mathematical modeling not only saves resources but also significantly shortens the drug development timeline. These technologies assist not only in the search for new molecules but also in choosing the optimal production method: instead of hundreds of experiments, researchers can calculate the most promising options and confirm them through a few experimental tests.
The head of Geropharm also expressed the opinion that in the future, mathematical modeling could be used at the proof stage — during clinical trials, where some data would be generated not only from experiments involving patients but also through computational models. “It’s an ambitious idea, but it needs to be discussed so that legal frameworks can be established later and the launch of innovative drugs in Russia and beyond can be accelerated,” he added.
After the pandemic, the pharmaceutical market began to decline: demand decreased, while regulatory attention increased. In such an environment, any advantages in speed, accuracy, and flexibility of development become strategically important. AI has become the key factor for acceleration and optimization. GxP News editorial analyzed how leading global pharmaceutical companies are integrating digital technologies into their strategies.