Seminar for
DECISION MAKING – THEORY, TECHNOLOGY AND PRACTICE

 

PROGRAM

---

Predavanja možete pratiti i online putem MITEAM stranice Seminara Odlučivanje - teorija, tehnologija, praksa:
https://miteam.mi.sanu.ac.rs/asset/sEL32w8mjmruyeEqW

---

Plan rada Seminara Odlučivanje - teorija, tehnologija, praksa za MAJ 2026.

---

Četvrtak, 14.05.2026. u 13:00, Online
Katarina Šmakić, Union Nikola Tesla University
ETHICS OF AUTONOMOUS DECISION-MAKING SYSTEMS
The presentation addresses the key issue of establishing responsibility in the context of the increasing autonomy of artificial intelligence systems, namely, whether responsibility can be attributed to machines or whether it ultimately remains in human hands. Drawing on contemporary philosophical and legal debates, the paper analyzes the boundaries between technical autonomy and moral responsibility, examining whether AI can be regarded as a bearer of moral agency or whether it remains an instrument of human decision-making. Particular attention is given to the problem of delegated decision-making, especially in high-risk domains such as autonomous weapon systems, where the consequences of decisions carry profound ethical and legal implications. The presentation explores different models of attributing responsibility, ranging from full human control, through shared responsibility, to the hypothetical moral autonomy of machines, and highlights the risks of prematurely shifting responsibility onto technology. In this context, it raises the question of whether the development of AI leads to a genuine redistribution of moral authority or rather reflects an attempt to evade human responsibility. In conclusion, the author argues that, despite the growing operational autonomy of AI systems, normative and legal responsibility must remain clearly anchored in human actors, along with the need for more precise regulatory frameworks to prevent the diffusion of responsibility within complex technological systems.

---

Četvrtak, 21.05.2026. u 13:00, Pariske Komune bb, Niš i Online
Dušan Tatić, Mathematical Institute of the Serbian Academy of Sciences and Arts
EXPERIMENTS IN CONSTRUCTING AI AVATARS FOR MUSEUM EXHIBITIONS
In this seminar, we will present the current stage of development of the museum AI avatar system. As part of the initial experimental phase of exploring AI technologies in museum environments, attention is given to the Retrieval-Augmented Generation (RAG) architecture. As a case study, we developed an AI avatar system based on an exhibition catalogue dedicated to the life and achievements of Mihailo Petrović Alas, which was prepared for the 150th anniversary of his birth.

---

Četvrtak, 28.05.2026. u 13:00, Online
Nilo Antonio de Souza Sampaio, State University of Rio de Janeiro, Brazil
APPLICATIONS OF STATISTICS AND ARTIFICIAL INTELLIGENCE IN ENVIRONMENTAL RESEARCH
The growing complexity of environmental problems requires analytical tools capable of dealing with large volumes of data, uncertainties and non-linear patterns. In this context, Statistics and Artificial Intelligence (AI) have played central roles in analysis, prediction and decision-making in environmental research. Statistics, traditionally used in environmental studies, allows for exploratory data analysis, hypothesis testing and probabilistic modeling. For example, statistical time series analysis has been widely applied to detect trends in climate data such as temperature and precipitation[5,6]. Statistical models are also essential for making inferences about water quality samples, species distribution and environmental risks[1]. On the other hand, Artificial Intelligence, especially Machine Learning techniques, has complemented Statistics by enabling the processing of large volumes of data with complex patterns. Artificial neural networks, random forests, and support vector machines (SVM) are commonly used in air quality forecasting, deforestation detection, and land use classification from satellite images[3]. The synergy between statistics and AI also allows for the development of hybrid models, combining statistical inference with the predictive power of AI algorithms, expanding.

---

---