From Explainable AI to Digital Twins: The NCUI Scientific Council Presented the Project’s Immediate Planned Outcomes

25.08.2026

The use of digital twins to optimize production and laboratory processes in healthcare, improved explainability of language and multimodal AI models, and anomaly detection software for assisted optimization in transportation. This is a summary of the first planned results of the National Center for Artificial Intelligence (NCUI), which the teams presented last week at a meeting of the NCUI Scientific Council. 

The leaders of the nine work packages presented the current status of the project’s solutions, planned outputs, and their future direction. The discussion focused not only on the upcoming outputs but also on their potential use in shaping public policies in the areas of artificial intelligence and digitalization, with an emphasis on the social implications and the ethical and legal aspects of AI. Linking cutting-edge research with the needs of society and industry is one of the key objectives of NCUI.

Four new software deliverables, developed in collaboration between research organizations and application partners, are expected to be completed by the end of the year. One of these is software for abductive explanations of decisions made by artificial intelligence methods, which is being developed in collaboration between the Faculty of Electrical Engineering (FEL) at the Czech Technical University in Prague (ČVUT) and GEN Digital (WP1). The goal is to enhance understanding of how artificial intelligence systems arrive at their decisions.

In the field of digital twins, two projects are currently underway (WP2). The first focuses on modeling and optimizing industrial processes through a hierarchical, modular digital twin of an industrial cyber-physical system. CIIRC CTU and Panasonic are collaborating on its development. 

The second project brings together CIIRC CTU, FIT CTU, Beckman Coulter, and Bulovka University Hospital to develop a digital twin of a laboratory system. Thus, all key stakeholders are involved—researchers, the hospital (and, by extension, the laboratory), and a technology integrator. The concept of a digital twin, widely used in manufacturing, will be applied in this case to optimize the entire sample flow process so that laboratory test results are available as quickly as possible. Using AI methods, the time to issue results (known as Turn-around Time – TAT) will be reduced, primarily in urgent cases (e.g., myocardial infarction). Another benefit is a reduction in clinical laboratory operating costs.

The fourth deliverable planned for this year is the AnomVision software, developed by the FEKT VUT team in collaboration with Yunex, a global market leader in intelligent traffic management (WP5). The goal is to detect anomalies in traffic videos, even in situations the system has not encountered during training. The result will be autonomous technology for AI-assisted optimization of intersection design, which—through the synergy of multi-agent systems and multimodal analysis of complex data streams—will contribute to the predictive and self-optimizing evolution of urban space.

The meeting confirmed that research teams and their industry partners are collaborating intensively on the development of technologies that have the potential to find applications in industry, healthcare, and transportation. NCUI thus continues to fulfill its mission—to translate artificial intelligence research results into practice and contribute to the development of innovations with real-world impact.