Medicines will be created 4 times faster: how AI has revolutionized pharmaceuticals

      From Giggles to Reality

      The development of artificial intelligence technologies is progressing at an astonishingly fast pace. As noted by Vadim Vitalievich Negrebetsky, the director of the Institute of Pharmacy and Medical Chemistry at Pirogov University, just a year ago many were skeptical about the prospects of AI in pharmaceuticals, but today "the fairy tale is becoming a reality."

      Previously, creating an original drug required at least 10-15 years and nearly a billion dollars, but now there are examples of entering clinical trials in just a year and a half. At the same time, the costs—both material and human—have become incomparably lower.

      Where AI is Already Helping

      The most obvious application is the development of new molecules. But the field is not limited to this. Artificial intelligence is actively used to identify patterns when working with big data: searching for reliable information about drugs, systematizing clinical data on efficacy and safety.

      Based on this data, decision support systems for physicians are developed. According to Vadim Vitalievich, AI allows doctors to "see the invisible." Analyzing millions of patient medical histories helps generate data for the development of innovative drugs. Personalized medicine is reaching a new level.

      Economics: A Fourfold Reduction in Costs and a Twofold Reduction in Time

      The pharmaceutical industry, as the expert explains, is not interested in creating "Aspirin_2"—a drug with only slightly greater efficacy but colossal development costs. The optimal path is to significantly increase the efficacy of drugs while simultaneously optimizing expenses.

      Drugs based on "small molecules" have likely reached the limit of their effectiveness. They may still be in demand in areas of unmet need: Parkinson's and Alzheimer's diseases, cognitive disorders, anxiety-depressive disorders, and recovery after a stroke.

      However, the main contribution of AI today is accelerating the initial stage of development, for example, "drawing" the initial lead molecule. All other stages (preclinical research and three phases of clinical trials for safety and efficacy) will still have to be passed "in vivo" by the drug candidate. But optimizing these stages, including clinical trial protocols, using machine learning and neural networks is quite feasible.

      According to available data, a reduction in investments in drug development through the application of AI is expected to be at least fourfold, and the time to market is expected to be reduced by at least half.

      Initial Results: 70 Drugs in the First Phase

      By the end of last year, about 70 drugs worldwide, created using AI, entered the first phase of clinical trials.

      This is a very good result; however, how many of them will pass safety and efficacy tests is still unclear. If we assume that the market entry percentage remains classic (around 12%), that is still not bad.

      A More Tempting Niche: Genetic Engineering

      According to the expert, an even more promising area for AI is not "small molecules," but genetic engineering methods. A recent example is the drug for Bechterew's disease "seniprutug" (an innovative development by Russian scientists), which is a recombinant human monoclonal antibody.

      Here, AI can help create effective treatments for previously incurable diseases.

      The Main Limitation: No Way Without Humans

      Vadim Vitalievich emphasizes: even in the future, AI will not be able to independently create a drug ready for market release.

      Without human involvement, everything that is governed by human consciousness will remain impossible: conducting experiments in the real world, verifying and interpreting the obtained data, conducting preclinical and clinical trials, regulatory studies, and much more.

      Artificial intelligence is already radically accelerating and reducing the costs of the initial stages of drug development—fourfold in costs and twofold in time. It helps doctors see the invisible and process vast amounts of data. But clinical trials on humans have not been canceled, and the final decision still rests with humans.