o:37649 University challenges: structural changes and quality of education Izazovi univerziteta: strukturne promene i kvalitet obrazovanja en ABSTRACT – The paper explores the challenges and problems that arise in the application of artificial intelligence and the latest ChatGPT technology at university. Recent discoveries in Natural Language Processing (NLP) have made it possible to synthesise and understand coherent text in an open way, translating theoretical algorithms into their practical application. The Large Language Model (LLM) has significantly improved the development of software for summarizing reports and authoring texts. However, recent empirical research has identified several ethical problems in the advanced ChatGPT model, highlighting its inherent shortcomings due to the ethical and social danger of irresponsible behaviour. These shortcomings include issues of compromising privacy, learning quality, transparency, and other ethical dimensions. To better understand and get rid of the practical aspects of ethical risks, it is necessary to do a system review and a user study on the ethics of how the ChatGPT model is currently used. In the meantime, the ethical aspects of using ChatGPT require careful planning and implementation, with clear guidelines on responsible use, in order to make the best possible use of the advantages of this technology while taking into account the risks and ethical dilemmas. Ovaj rad istražuje izazove i probleme koji nastaju u primeni veštačke inteligencije i najnovije ChatGPT tehnologije na univerzitetu. Nedavna otkrića u obradi prirodnog jezika (eng. Natural Language Processing - NLP) omogućila su sintezu i razumevanje koherentnog teksta na otvoren način, prevodeći teorijske algoritme u njihovu praktičnu primenu. Veliki jezički model (eng. Large Language Model - LLM) značajno je unapredio izradu softvera za sumiranje izveštaja i autorizovanje tekstova. Međutim, nedavna empirijska istraživanja su utvrdila nekoliko etičkih problema u naprednom ChatGPT modelu i pokazalo se da ovaj model veštačke inteligencije ima svoje nedostatke u vidu etičke i društvene opasnosti od neodgovornog ponašanja koje uključuje pitanja ugrožavanja privatnosti, kvaliteta učenja, transparentnosti i druge etičke aspekte. Da bi se to ispravilo potrebno je uraditi sistemsko ispitivanje i korisničku studiju o etičnosti trenutne upotrebe ChatGPT modela kako bi se bolje razumele i otklonile praktične karakteristike etičkih opasnosti. U medjuvremenu, etički aspekti korišćenja ChatGPT-a zahtevaju njegovo pažljivo planiranje i implementaciju, uz jasne smernice o odgovornom korišćenju, kako bi se na najbolji način iskoristile prednosti ove tehnologije, uz vođenje računa o rizicima i etičkim dilemama. KEYWORDS: artificial intelligence, ethical issues, shortcomings, dangers KLJUČNE REČI: strukturne promene, kvalitet visokog obrazovanja 1552100 978-961-6770-79-8 2026-01-27T14:23:49.641Z 44 yes 46 Aleksandar Stojanović University of Belgrade, Faculty of Education, Belgrade 0000-0002-0060-1194 Grozdanka Gojkov Serbian Academy of Education person application/pdf 359626 https://phaidrabg.bg.ac.rs/o:37649 no yes 1 70 1556235 11A42 Globalisation Challenges And The Social Economic Environment Of The EU 494 506 University of Novo mesto Faculty of Economics and Informatics 2024