2026-05-09T17:49:43.966Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:1:11455 2025-06-18T21:22:33Z 1

Ethical Concerns and Mass Agentic AI Adoption 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/1/11455 https://portal.sinteza.singidunum.ac.rs/paper/1059 S. Jovanović L. Dražeta A. Petrović M. Živković N. Bačanin Džakula The arrival of agentic artificial intelligence (AI) marks a significant jump in the evolution of intelligent systems. Unlike AI so far, which primarily operates within predefined parameters, agentic AI possesses the capability to autonomously perceive, reason, act, and learn from its environment. This new paradigm enables AI agents to solve complex, multi–step problems and interact with various tools and data sources independently. Rather than executing commands or routines set by humans, Agentic AI systems are intended to adjust their strategies and explore their environments as needed. Agentic AI's sophisticated reasoning and iterative planning abilities are settled to transform numerous industries. From optimizing supply chains and personalizing customer service to enhancing healthcare and automating complex tasks, the potential applications are vast and varied. However, with these advancements come challenges, including the need to address biases, ensure ethical use, and establish robust regulatory frameworks. This paper explores the foundational principles of Agentic AI, its current and potential applications, and the ethical considerations that must guide its development and deployment. By examining these aspects, the aim is to provide a good understanding of how Agentic AI can be used to drive innovation while mitigating associated risks. conferenceObject 369 375 10.15308/Sinteza-2025-369-375 Sinteza 2025 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Serbia