Evaluating science: A comparison of human and AI reviewers

Scientists have started to explore whether novel artificial intelligence (AI) tools based on large language models, such as GPT-4, could support the scientific peer review process. We sought to understand (i) whether AI versus human reviewers are able to distinguish between made-up AI-generated and...

Full description

Saved in:
Bibliographic Details
Main Authors: Anna Shcherbiak, Hooman Habibnia, Robert Böhm, Susann Fiedler
Format: Article
Language:English
Published: Cambridge University Press 2024-01-01
Series:Judgment and Decision Making
Subjects:
Online Access:https://www.cambridge.org/core/product/identifier/S193029752400024X/type/journal_article
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846161412044357632
author Anna Shcherbiak
Hooman Habibnia
Robert Böhm
Susann Fiedler
author_facet Anna Shcherbiak
Hooman Habibnia
Robert Böhm
Susann Fiedler
author_sort Anna Shcherbiak
collection DOAJ
description Scientists have started to explore whether novel artificial intelligence (AI) tools based on large language models, such as GPT-4, could support the scientific peer review process. We sought to understand (i) whether AI versus human reviewers are able to distinguish between made-up AI-generated and human-written conference abstracts reporting on actual research, and (ii) how the quality assessments by AI versus human reviewers of the reported research correspond to each other. We conducted a large-scale field experiment during a medium-sized scientific conference, relying on 305 human-written and 20 AI-written abstracts that were reviewed either by AI or 217 human reviewers. The results show that human reviewers and GPTZero were better in discerning (AI vs. human) authorship than GPT-4. Regarding quality assessments, there was rather low agreement between both human–human and human–AI reviewer pairs, but AI reviewers were more aligned with human reviewers in classifying the very best abstracts. This indicates that AI could become a prescreening tool for scientific abstracts. The results are discussed with regard to the future development and use of AI tools during the scientific peer review process.
format Article
id doaj-art-77b7e2b6ba6046d9915c704b5d20d1be
institution Kabale University
issn 1930-2975
language English
publishDate 2024-01-01
publisher Cambridge University Press
record_format Article
series Judgment and Decision Making
spelling doaj-art-77b7e2b6ba6046d9915c704b5d20d1be2024-11-21T07:18:15ZengCambridge University PressJudgment and Decision Making1930-29752024-01-011910.1017/jdm.2024.24Evaluating science: A comparison of human and AI reviewersAnna Shcherbiak0Hooman Habibnia1https://orcid.org/0009-0000-3915-2624Robert Böhm2Susann Fiedler3Institute for Cognition and Behavior, WU Vienna University of Economics and Business, Vienna, AustriaInstitute for Cognition and Behavior, WU Vienna University of Economics and Business, Vienna, AustriaFaculty of Psychology, University of Vienna, Vienna, Austria Department of Psychology and Copenhagen Center for Social Data Science (SODAS), University of Copenhagen, Copenhagen, DenmarkInstitute for Cognition and Behavior, WU Vienna University of Economics and Business, Vienna, AustriaScientists have started to explore whether novel artificial intelligence (AI) tools based on large language models, such as GPT-4, could support the scientific peer review process. We sought to understand (i) whether AI versus human reviewers are able to distinguish between made-up AI-generated and human-written conference abstracts reporting on actual research, and (ii) how the quality assessments by AI versus human reviewers of the reported research correspond to each other. We conducted a large-scale field experiment during a medium-sized scientific conference, relying on 305 human-written and 20 AI-written abstracts that were reviewed either by AI or 217 human reviewers. The results show that human reviewers and GPTZero were better in discerning (AI vs. human) authorship than GPT-4. Regarding quality assessments, there was rather low agreement between both human–human and human–AI reviewer pairs, but AI reviewers were more aligned with human reviewers in classifying the very best abstracts. This indicates that AI could become a prescreening tool for scientific abstracts. The results are discussed with regard to the future development and use of AI tools during the scientific peer review process.https://www.cambridge.org/core/product/identifier/S193029752400024X/type/journal_articleartificial intelligencelarge language modelspeer reviewquality assessment
spellingShingle Anna Shcherbiak
Hooman Habibnia
Robert Böhm
Susann Fiedler
Evaluating science: A comparison of human and AI reviewers
Judgment and Decision Making
artificial intelligence
large language models
peer review
quality assessment
title Evaluating science: A comparison of human and AI reviewers
title_full Evaluating science: A comparison of human and AI reviewers
title_fullStr Evaluating science: A comparison of human and AI reviewers
title_full_unstemmed Evaluating science: A comparison of human and AI reviewers
title_short Evaluating science: A comparison of human and AI reviewers
title_sort evaluating science a comparison of human and ai reviewers
topic artificial intelligence
large language models
peer review
quality assessment
url https://www.cambridge.org/core/product/identifier/S193029752400024X/type/journal_article
work_keys_str_mv AT annashcherbiak evaluatingscienceacomparisonofhumanandaireviewers
AT hoomanhabibnia evaluatingscienceacomparisonofhumanandaireviewers
AT robertbohm evaluatingscienceacomparisonofhumanandaireviewers
AT susannfiedler evaluatingscienceacomparisonofhumanandaireviewers