From Fundamental Science to Saving Lives: How AI is Developing in Belarus
15.07.2026
On the eve of Artificial Intelligence Day (July 16), a press conference entitled "Belarus' Scientific Potential and Developments in AI: Trends, Challenges, and Intellectual Property Protection" was held in Minsk. Representatives of the United Institute of Informatics Problems of the National Academy of Sciences of Belarus presented a comprehensive view of technological developments—from the creation of neural networks to save lives to the drafting of laws that will shape the future of the country's digital space. The event was attended by Sergey Kasanin, Deputy Director General for Research, Eduard Snezhko, Head of the Biomedical Image Analysis Laboratory, and Anna Karpenko, Research Fellow at the Laboratory of Mathematical Cybernetics.
Legal Foundation
Artificial intelligence is developing rapidly, bringing not only new opportunities but also serious risks. The issue of legal regulation of AI is currently a pressing issue. As Sergey Kasanin noted, the Republic of Belarus is in the process of forming a national legislative framework, but the scientific community has already proposed a ready-made architecture for the future law. "We understand that we need some specific legislation that would regulate artificial intelligence technologies. And we are approaching this," he emphasized.
Back in 2023, the UIPI won an open competition held by the CIS Interparliamentary Assembly to develop a model law. The result of this extensive work was a document adopted in April 2025 – the Law "On Artificial Intelligence Technologies." The document describes the basic concepts and principles of work in the field of AI.
Medicine without Borders
The Biomedical Image Analysis Laboratory, led by Eduard Vitalyevich Snezhko, has a unique background – it has been working with medical images for over 30 years. The results of their work are impressive: in a number of highly specialized tasks, artificial intelligence is already helping doctors identify pathologies in X-rays and tomograms at the earliest stages.
The laboratory's key research areas cover the full cycle of security and implementation:
Assistance to doctors: scientists actively collaborate with traumatologists, cardiologists, oncologists, and pulmonologists.
AI security: developing methods to protect neural networks from malicious attacks, scientists are exploring scenarios where input changes, almost imperceptible to the human eye, cause the system to generate a fatal error.
Federated learning is a technology that allows AI to be trained on data without sharing it outside hospitals and clinics, solving the problem of medical confidentiality and cybersecurity of medical databases.
Leak protection: finding ways to prevent the recovery of patients' personal data from the parameters of an already trained neural network.
Digital alchemical art
Anna Dmitrievna Karpenko spoke in detail about one of the key areas of work at the Mathematical Cybernetics Laboratory: drug design. She focused on three main areas of focus:
Repurposing medicinal pharmaceutical compounds (in collaboration with a research group from China's Fudan University, Belarusian scientists were able to validate computational models that enabled the identification of an effective compound against SARS-CoV-2; the results of this work are currently undergoing patent protection);
Identification of molecular targets: understanding the spatial structure of a target protein is a fundamental step in rational drug design. Structural modeling approaches developed in the laboratory allow for the visualization of active sites and the prediction of probable ligand binding sites, which is critical for reducing the risks of late-stage clinical trials. This approach will further reduce the risk of side effects and identify new therapeutic targets;
Generative design (de novo molecular design): Unlike classical methods of virtual screening of existing databases, generative neural network architectures make it possible to construct new chemical structures with specified pharmacokinetic properties, expanding the chemical search space beyond known compound libraries and generating new drugs. Compounds against cancer, SARS-CoV-2, and tuberculosis are currently being tested by partners from the Chinese Academy of Sciences, Fudan University, and the National Academy of Sciences of Belarus.
The topic of artificial intelligence is multifaceted and complex; this phenomenon is actively developing and is only just becoming integrated into the familiar world order.
Legal Foundation
Artificial intelligence is developing rapidly, bringing not only new opportunities but also serious risks. The issue of legal regulation of AI is currently a pressing issue. As Sergey Kasanin noted, the Republic of Belarus is in the process of forming a national legislative framework, but the scientific community has already proposed a ready-made architecture for the future law. "We understand that we need some specific legislation that would regulate artificial intelligence technologies. And we are approaching this," he emphasized.
Back in 2023, the UIPI won an open competition held by the CIS Interparliamentary Assembly to develop a model law. The result of this extensive work was a document adopted in April 2025 – the Law "On Artificial Intelligence Technologies." The document describes the basic concepts and principles of work in the field of AI.
Medicine without Borders
The Biomedical Image Analysis Laboratory, led by Eduard Vitalyevich Snezhko, has a unique background – it has been working with medical images for over 30 years. The results of their work are impressive: in a number of highly specialized tasks, artificial intelligence is already helping doctors identify pathologies in X-rays and tomograms at the earliest stages.
The laboratory's key research areas cover the full cycle of security and implementation:
Assistance to doctors: scientists actively collaborate with traumatologists, cardiologists, oncologists, and pulmonologists.
AI security: developing methods to protect neural networks from malicious attacks, scientists are exploring scenarios where input changes, almost imperceptible to the human eye, cause the system to generate a fatal error.
Federated learning is a technology that allows AI to be trained on data without sharing it outside hospitals and clinics, solving the problem of medical confidentiality and cybersecurity of medical databases.
Leak protection: finding ways to prevent the recovery of patients' personal data from the parameters of an already trained neural network.
Digital alchemical art
Anna Dmitrievna Karpenko spoke in detail about one of the key areas of work at the Mathematical Cybernetics Laboratory: drug design. She focused on three main areas of focus:
Repurposing medicinal pharmaceutical compounds (in collaboration with a research group from China's Fudan University, Belarusian scientists were able to validate computational models that enabled the identification of an effective compound against SARS-CoV-2; the results of this work are currently undergoing patent protection);
Identification of molecular targets: understanding the spatial structure of a target protein is a fundamental step in rational drug design. Structural modeling approaches developed in the laboratory allow for the visualization of active sites and the prediction of probable ligand binding sites, which is critical for reducing the risks of late-stage clinical trials. This approach will further reduce the risk of side effects and identify new therapeutic targets;
Generative design (de novo molecular design): Unlike classical methods of virtual screening of existing databases, generative neural network architectures make it possible to construct new chemical structures with specified pharmacokinetic properties, expanding the chemical search space beyond known compound libraries and generating new drugs. Compounds against cancer, SARS-CoV-2, and tuberculosis are currently being tested by partners from the Chinese Academy of Sciences, Fudan University, and the National Academy of Sciences of Belarus.
The topic of artificial intelligence is multifaceted and complex; this phenomenon is actively developing and is only just becoming integrated into the familiar world order.