20.500.12556/RUL-182059 Razvoj in delovanje umetne inteligence V tem besedilu se bomo na seznanili z razvojem in značilnostmi umetne inteligence (UI), ki je postala del našega vsakdanjika. Osredotočili se bomo na generativno UI, kot so ChatGPT in DALL-E, ki spontano prodirajo v izobraževanje in prinašajo priložnosti za individualizacijo učenja ter različne izzive. V prvem delu spoznamo osnove UI – sisteme, ki se učijo iz podatkov. Razlikujemo med ozko UI (današnji specializirani sistemi) in splošno UI (še ne obstaja). Sledil bo zgodovinski pregled od Turingovega testa (1950) do sodobnih transformatorjev. Nadaljevali bomo z generativno UI, ki ustvarja nove vsebine in deluje večmodalno. Razložili bomo, kako veliki jezikovni modeli (LLM) statistično napovedujejo besede, ne razumejo pa sveta vzročno, kar povzroča halucinacije. Opozorili bomo, da UI ni nepristranska in odraža pristranskosti učnih podatkov. Zaključili bomo s ključno nevarnostjo – kognitivnim dolgom, kjer prekomerna odvisnost od UI oslabi kritično mišljenje. Poudarili bomo, da mora UI ostati orodje za razširitev razmišljanja, učitelj pa nepogrešljiv vodja učnega procesa. In this text, we will become acquainted with the development and characteristics of artificial intelligence (AI), which has become part of our everyday life. We will focus on generative AI, such as ChatGPT and DALL-E, which are spontaneously penetrating education and bringing opportunities for individualisation of learning as well as various chal-lenges. In the first part, we will explore the basics of AI – systems that learn from data. We distinguish between narrow AI (today‘s specialised systems) and general AI (does not yet exist). This will be followed by a historical overview from the Turing test (1950) to contemporary transformers. We will continue with generative AI, which creates new content and operates multimodally. We will explain how large language models (LLMs) statistically predict words but do not understand the world causally, which causes hal-lucinations. We will point out that AI is not unbiased and reflects the biases of training data. We will conclude with a key danger – cognitive debt, where excessive reliance on AI weakens critical thinking. We will emphasise that AI must remain a tool for expanding thinking, whilst the teacher remains an indispensable leader of the learning process. generativna umetna inteligenca (Gen-UI) kognitivni dolg pedagoška nadgradnja algoritemska pristranskost generative artificial intelligence (Gen-AI) cognitive debt pedagogical enhancement algorithmic bias true false true Slovenski jezik Angleški jezik Neznano 2026-04-23 11:12:23 2026-04-24 11:34:07 2026-04-25 04:48:40 0000-00-00 00:00:00 2026 0 0 Str. 9-19 0000-00-00 Zaloznikova Objavljeno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 159.95 276097027 274395651 Radovan,+Gradisek.pdf Radovan,+Gradisek.pdf 1 6CDDA599432A71709E61C2136E27AFC4 1b98fe32d0329825dda31552475b0135982dbda0c59a9b214ca454c41765064a 76921d6f-3ef4-11f1-b0ab-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=232987 https://ebooks.uni-lj.si/ZalozbaUL/catalog/view/911/1305/13110 1 7dce9dd3-3fc0-11f1-b0ab-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=233055 Filozofska fakulteta 0 0 0