20.500.12556/DKUM-95350 Stanje stereotipov o podobi računovodij: analiza odgovorov umetne inteligence The state of stereotypes about the image of accountants: an analysis of artificial intelligence responses Namen diplomskega dela je bil raziskati različne stereotipe o računovodjih, poklicu računovodje ter o njihovem delu. Poudarek je bil na vprašanju, kako različne aplikacije umetne inteligence (UI) te stereotipe razumejo oziroma kakšna imajo stališča glede posameznega stereotipa. Diplomsko delo je osredotočeno na analizo odgovorov petih različnih aplikacij UI. V uvodnem delu diplomskega dela je naprej kratka predstavitev računovodje z opredelitvijo njegovih značilnosti ter njegovega poklica oziroma dela. V tretjem poglavju je predstavljeno, kaj so stereotipi ter kateri so stereotipi o računovodjih. V četrtem poglavju je na kratko predstavljena UI ter aplikacije UI, katere so vključene v diplomsko delo. Te aplikacije so ChatGPT, Copilot, DeepSeek, Claude AI ter Gemini. V tem poglavju pa so v okviru raziskave analizirani odgovori teh aplikacij UI na vprašanja, ki so o obstoječih stereotipih o računovodjih. Rezultati raziskave prikazujejo, da UI pogosto zavračajo obstoječe stereotipe, vendar pa jih v nekaterih primerih tudi potrdijo. V slednjih primerih pa so jim predstavljena nasprotna dejstva, kar pa vpliva na to, da aplikacije UI svoja stališča spremenijo. To dokazuje, da je na odgovore UI s predstavitvijo konkretnih dejstev možno vplivati. V zaključku so obravnavane hipoteze. Prva hipoteza, »vse aplikacije UI imajo enaka stališča o računovodjih in njihovem delu« je bila zavrnjena, saj aplikacije pri svojih odgovorih niso imele popolnoma enakih stališč, prav tako pa so nekatere aplikacije ob predstavitvi dodatnih dejstev svoje stališče spremenile. Tudi druga hipoteza »vse aplikacije UI v svojih odgovorih potrjujejo obstoječe stereotipe o računovodjih«, je bila zavrnjena, saj aplikacije UI vseh stereotipov niso potrdile. Analiza je sicer pokazala, da nekatere stereotipe potrjujejo, na primer, »računovodje obožujejo preglednice«, mnoge pa zavračajo. Tretja hipoteza »s predstavitvijo nasprotnih dejstev je možno posamezno aplikacijo UI prepričati v spremembo svojega stališča« je bila sprejeta, saj je bilo pri več primerih možno spremeniti oziroma vplivati na stališče aplikacije UI. V raziskavi je bilo ugotovljeno, da UI ne predstavlja enotnega vira informacij, temveč deluje kot sistem, ki ga je možno prilagoditi oziroma mu spremeniti stališče z predstavitvijo konkretnih dejstev. V prihodnje bi bilo smiselno raziskati tudi vpliv UI na družbene stereotipe o drugih poklicih. The bachelor's degree thesis aimed to investigate various stereotypes about accountants, the accounting profession and their work. The emphasis was on how different artificial intelligence (AI) applications understand these stereotypes and their views regarding each stereotype. The thesis focuses on the analysis of responses from five different AI applications. The introductory part of the thesis briefly introduces the accountant with a definition of his characteristics and his profession or work. The third chapter presents what stereotypes are and what stereotypes about accountants. The fourth chapter briefly presents AI and the AI applications included in the thesis. These applications are ChatGPT, Copilot, DeepSeek, Claude AI, and Gemini. In this chapter, the research analyses the responses of these AI applications to questions about existing stereotypes about accountants. The results of the research show that AIs often rejects existing stereotypes, but in some cases, they also confirm them. In the latter cases, however, they are presented with opposing facts, which in turn causes the AI applications to change their positions. This proves that AI responses can be influenced by presenting concrete facts. The hypotheses are discussed in the conclusion. The first hypothesis, that all AI applications have the same views on accountants and their work, was rejected because the applications did not have completely identical views in their responses, and some applications changed their views when presented with additional facts. The second hypothesis, that all AI applications confirm existing stereotypes about accountants in their responses, was also rejected, as the AI applications did not confirm all stereotypes. The analysis showed that some stereotypes are confirmed, for example, that accountants love spreadsheets, but many are rejected. The third hypothesis, that by presenting counterarguments, it is possible to convince an individual AI application to change its position, was accepted, as in several cases it was possible to change or influence the position of the AI application. The study found that AI does not represent a single source of information, but functions as a system that can be adapted or whose position can be changed by presenting concrete facts. In the future, it would also be helpful to investigate the influence of AI on social stereotypes about other professions. Računovodje stereotipi generativna umetna inteligenca analiza odgovorov UI sprememba stereotipa spreminjanje mnenja UI Accountants stereotypes generative artificial intelligence analysis of AI responses changing stereotype changing the mind of AI true true false N. Pajtler] Slovenski jezik Angleški jezik Diplomsko delo/naloga 2025-09-15 10:42:48 2025-10-27 08:52:51 2025-10-28 03:24:12 0000-00-00 00:00:00 2025 0 Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 657:004.8 254811139 UN_Pajtler_Nastja_2025.pdf UN_Pajtler_Nastja_2025.pdf 1 080C43F64888F89A5975B267D1184ACB a75c5c1b228e40f730ca8c88f6453ca67e2137f2cf75c5a4d7df725f8be39474 72ac6863-97f7-11f0-88a2-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=198998 Ekonomsko-poslovna fakulteta 0 0 0