Can we identify prominent scholars using ChatGPT?

dc.contributor.author	Sandnes, Frode Eika
dc.date.accessioned	2023-12-14T14:59:44Z
dc.date.available	2023-12-14T14:59:44Z
dc.date.created	2023-12-02T16:00:49Z
dc.date.issued	2023
dc.identifier.citation	Scientometrics. 2023, .	en_US
dc.identifier.issn	0138-9130
dc.identifier.issn	1588-2861
dc.identifier.uri	https://hdl.handle.net/11250/3107660
dc.description.abstract	It may be tempting to learn about scholars using ChatGPT. To validate ChatGPT for this task a small experiment was conducted based on the 50 most cited researchers at the author’s university. The results show that ChatGPT (GPT-3.5) only recognized an ad-hoc subset of the scholars with no obvious connection to the respective authors’ citation attributes or internet footprint. Moreover, details about the recognized scholars were often erroneous.	en_US
dc.language.iso	eng	en_US
dc.publisher	Springer	en_US
dc.relation.ispartofseries	Scientometrics;
dc.title	Can we identify prominent scholars using ChatGPT?	en_US
dc.type	Peer reviewed	en_US
dc.type	Journal article	en_US
dc.description.version	acceptedVersion	en_US
cristin.ispublished	true
cristin.fulltext	original
cristin.fulltext	postprint
cristin.qualitycode	1
dc.identifier.doi	https://doi.org/10.1007/s11192-023-04882-4
dc.identifier.cristin	2207775
dc.source.journal	Scientometrics	en_US
dc.source.pagenumber	6	en_US