73 KB601 KB36 KB2025Hoeren, ThomasLajovic, Tomaž1 spletni vir (1 datoteka PDF (12, II str.))COBISSID:252408067ISBN:978-961-07-2932-7URN:URN:NBN:SI:doc-4KFY03QMenSplatoknjigeAvtorske praviceavtorsko pravobaze podatkovelektronske knjigeUmetna inteligencaFoundation models as infringers: should large-scale AI training trigger collective licensing obligations under EU law?| legal essay as part of the certificate course Law of the Digital Economy|The essay by Tomaž Lajovic addresses the copyright implications of foundation models (FMs), particularly large language models, under EU law. It examines how training datasets, memorisation, and output generation implicate copyright and database rights, and assesses whether collective licensing regimes or exemptions with remuneration rights could balance the interests of AI developers and rightsholdersEsej Tomaža Lajovca obravnava avtorskopravne posledice temeljnih modelov umetne inteligence (ang. "Foundation Model", FM) - zlasti velikih jezikovnih modelov - v skladu z zakonodajo EU. Preučuje, kako nabori učnih podatkov (ang. "Learning dataset"), pomnjenje (ang. "Memorisation") in ustvarjanje rezultatov z umetno inteligenco vplivajo na avtorske pravice in sui generis pravice na bazah podatkov, ter ocenjuje, ali bi režim kolektivnega licenciranja oziroma omejitev avtorske pravice s pravico do nadomestila lahko uravnotežili interese razvijalcev umetne inteligence in imetnikov pravicTEXTknjigebooksSlovenian National E-content AggregatorNational and University Library of Slovenia