Logo des Repositoriums
 

Auto-generated language learning online courses using generative AI models like ChatGPT

dc.contributor.authorRüdian, Sylvio
dc.contributor.authorPinkwart, Niels
dc.contributor.editorRöpke, René
dc.contributor.editorSchroeder, Ulrik
dc.date.accessioned2023-08-30T09:09:41Z
dc.date.available2023-08-30T09:09:41Z
dc.date.issued2023
dc.description.abstractGenerating online courses is always a trade-off between possibilities, technical limitations, and quality. State-of-the-art generative models can assist teachers in the creation process. However, generating learning materials is highly complex. Hence, teachers mainly create them manually. In this paper, learning content for a concrete micro-learning template is generated focusing on the field of language teaching. It intends that learners can find correct responses by logical thinking. Teachers provide a topic as input. Then, the approach asks for the required information using GPT3.5 with instructional prompts and combines responses to form a language learning unit. The quality of the resulting learning content, focusing on correctness, and appropriateness, is evaluated and discussed to examine the practicability of the tool, and alternatives are given.en
dc.identifier.doi10.18420/delfi2023-14
dc.identifier.isbn978-3-88579-732-6
dc.identifier.pissn1617-5468
dc.identifier.urihttps://dl.gi.de/handle/20.500.12116/42239
dc.language.isoen
dc.publisherGesellschaft für Informatik e.V.
dc.relation.ispartof21. Fachtagung Bildungstechnologien (DELFI)
dc.relation.ispartofseriesLecture Notes in Informatics (LNI) - Proceedings, Volume P-322
dc.subjectLanguage learning
dc.subjectgenerative AI models
dc.subjectauto-generated course units
dc.titleAuto-generated language learning online courses using generative AI models like ChatGPTen
dc.typeText/Conference Paper
gi.citation.endPage76
gi.citation.publisherPlaceBonn
gi.citation.startPage65
gi.conference.date11.-13. September 2023
gi.conference.locationAachen
gi.conference.reviewfull
gi.conference.sessiontitleBest-Paper-Kandidaten

Dateien

Originalbündel
1 - 1 von 1
Vorschaubild nicht verfügbar
Name:
14.pdf
Größe:
842.57 KB
Format:
Adobe Portable Document Format