20.500.12556/DKUM-95327 Znanje bodočih učiteljev razrednega pouka o pisanju pozivov v klepetalnikih magistrsko delo Future primary teachers' knowledge about writing prompts in chatbots Uporaba umetne inteligence je v zadnjih nekaj letih sunkovito narasla. Integrira se v vsa področja našega življenja kot tudi v vzgojno-izobraževalno panogo. Pomembno področje je generativna UI, natančneje veliki jezikovni modeli, na podlagi katerih delujejo klepetalniki, kot je npr. ChatGPT. Predhodne raziskave kažejo, da je znanje študentov o pisanju pozivov v klepetalnikih šibko. Namen magistrskega dela je bil ugotoviti, kako izobraževanje o tehnikah in metodah pisanja pozivov vpliva na znanje bodočih učiteljev razrednega pouka. V raziskavi je sodelovalo 82 študentov razrednega pouka. Pozive smo zbrali pred in po izobraževanju in jih obdelali s kvalitativno vsebinsko analizo. Potrdili smo, da študenti ne obvladajo dovolj veščin za pisanje pozivov in da je ciljno izobraževanje potrebno in koristno. Rezultati kažejo, da so po izobraževanju študenti oblikovali bolj kakovostne pozive kot pred izobraževanjem. Krepko se je povečal delež osnovnih strategij, še vedno pa so imeli težave pri zahtevnejših strategijah, za katere predlagamo dodatna izobraževanja. The use of artificial intelligence has grown rapidly in recent years. It is being integrated into all areas of our lives, including education. An important area is generative AI, specifically large language models, on which chatbots such as ChatGPT are based. Previous research shows that students' knowledge of writing prompts in chatbots is limited. The purpose of this master's thesis was to determine how training in prompt engineering techniques and methods affects the knowledge of future primary teachers. The study included 82 students of primary education. We collected prompts before and after the training and analyzed them using qualitative content analysis. The findings confirmed that students lack sufficient skills for writing effective prompts and that targeted training is necessary and beneficial. The results show that after the training, students created higher quality prompts. The percentage of basic strategies increased significantly, but students continued to struggle with more complex strategies, for which we suggest additional training. klepetalnik pisanje pozivov generativna umetna inteligenca chatbot prompt engineering generative artificial intelligence true true false V. Doler Slovenski jezik Angleški jezik Magistrsko delo/naloga 2025-09-13 10:09:05 2025-10-17 09:28:02 2025-10-18 03:21:12 0000-00-00 00:00:00 2025 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF (XIII, 84 str.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.8:37-057.875(043.2) 253598723 MAG_Doler_Vita_2025.pdf MAG_Doler_Vita_2025.pdf 1 0FFE4D65639669BE372A6C1200C1BEA5 63a9ef8c2ae2742d434fd4a2a2f79ac9a6b17b8846af4b85c7eb1115a1edca0a c94da783-9088-11f0-9526-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=198339 Pedagoška fakulteta 0 0 0