ChatGPT outperforms crowd workers for text-annotation tasks

Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models. Depending on the size and degree of complexity, the tasks may be conducted by crowd workers on platforms such as MTurk as well as trained ann...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings of the National Academy of Sciences - PNAS Jg. 120; H. 30; S. e2305016120
Hauptverfasser: Gilardi, Fabrizio, Alizadeh, Meysam, Kubli, Maël
Format: Journal Article
Sprache:Englisch
Veröffentlicht: United States 25.07.2023
Schlagworte:
ISSN:1091-6490, 1091-6490
Online-Zugang:Weitere Angaben
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models. Depending on the size and degree of complexity, the tasks may be conducted by crowd workers on platforms such as MTurk as well as trained annotators, such as research assistants. Using four samples of tweets and news articles ( = 6,183), we show that ChatGPT outperforms crowd workers for several annotation tasks, including relevance, stance, topics, and frame detection. Across the four datasets, the zero-shot accuracy of ChatGPT exceeds that of crowd workers by about 25 percentage points on average, while ChatGPT's intercoder agreement exceeds that of both crowd workers and trained annotators for all tasks. Moreover, the per-annotation cost of ChatGPT is less than $0.003-about thirty times cheaper than MTurk. These results demonstrate the potential of large language models to drastically increase the efficiency of text classification.
AbstractList Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models. Depending on the size and degree of complexity, the tasks may be conducted by crowd workers on platforms such as MTurk as well as trained annotators, such as research assistants. Using four samples of tweets and news articles (n = 6,183), we show that ChatGPT outperforms crowd workers for several annotation tasks, including relevance, stance, topics, and frame detection. Across the four datasets, the zero-shot accuracy of ChatGPT exceeds that of crowd workers by about 25 percentage points on average, while ChatGPT's intercoder agreement exceeds that of both crowd workers and trained annotators for all tasks. Moreover, the per-annotation cost of ChatGPT is less than $0.003-about thirty times cheaper than MTurk. These results demonstrate the potential of large language models to drastically increase the efficiency of text classification.Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models. Depending on the size and degree of complexity, the tasks may be conducted by crowd workers on platforms such as MTurk as well as trained annotators, such as research assistants. Using four samples of tweets and news articles (n = 6,183), we show that ChatGPT outperforms crowd workers for several annotation tasks, including relevance, stance, topics, and frame detection. Across the four datasets, the zero-shot accuracy of ChatGPT exceeds that of crowd workers by about 25 percentage points on average, while ChatGPT's intercoder agreement exceeds that of both crowd workers and trained annotators for all tasks. Moreover, the per-annotation cost of ChatGPT is less than $0.003-about thirty times cheaper than MTurk. These results demonstrate the potential of large language models to drastically increase the efficiency of text classification.
Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models. Depending on the size and degree of complexity, the tasks may be conducted by crowd workers on platforms such as MTurk as well as trained annotators, such as research assistants. Using four samples of tweets and news articles ( = 6,183), we show that ChatGPT outperforms crowd workers for several annotation tasks, including relevance, stance, topics, and frame detection. Across the four datasets, the zero-shot accuracy of ChatGPT exceeds that of crowd workers by about 25 percentage points on average, while ChatGPT's intercoder agreement exceeds that of both crowd workers and trained annotators for all tasks. Moreover, the per-annotation cost of ChatGPT is less than $0.003-about thirty times cheaper than MTurk. These results demonstrate the potential of large language models to drastically increase the efficiency of text classification.
Author Gilardi, Fabrizio
Alizadeh, Meysam
Kubli, Maël
Author_xml – sequence: 1
  givenname: Fabrizio
  orcidid: 0000-0002-0635-3048
  surname: Gilardi
  fullname: Gilardi, Fabrizio
  organization: Department of Political Science, University of Zurich, Zurich 8050, Switzerland
– sequence: 2
  givenname: Meysam
  orcidid: 0000-0001-6696-6471
  surname: Alizadeh
  fullname: Alizadeh, Meysam
  organization: Department of Political Science, University of Zurich, Zurich 8050, Switzerland
– sequence: 3
  givenname: Maël
  orcidid: 0000-0002-5592-9648
  surname: Kubli
  fullname: Kubli, Maël
  organization: Department of Political Science, University of Zurich, Zurich 8050, Switzerland
BackLink https://www.ncbi.nlm.nih.gov/pubmed/37463210$$D View this record in MEDLINE/PubMed
BookMark eNpNj81Lw0AUxBep2A89e5McvaS-fbvZJHiSUqtQ0EM9h03yFmub3bi7ofrfW7CCpxmGH8PMlI2ss8TYNYc5h1zc9VaHOQrIgCuOcMYmHEqeKlnC6J8fs2kIHwBQZgVcsLHIpRLIYcLuF-86rl43iRtiT94434Wk8e7QJgfnd-RDcsySSF8x1da6qOPW2STqsAuX7NzofaCrk87Y2-Nys3hK1y-r58XDOm0ylDHlQooW5XElJ8OxAFWQQiOUIYOyRaU1QC1zKgFkA3ltsCgaiVmWQc3bGmfs9re39-5zoBCrbhsa2u-1JTeECgtR5pJLUR7RmxM61B21Ve-3nfbf1d9h_AHYoFio
CitedBy_id crossref_primary_10_1002_jee_70033
crossref_primary_10_1017_pan_2025_10017
crossref_primary_10_1038_s41598_025_92002_y
crossref_primary_10_3390_foods14081355
crossref_primary_10_1177_20563051241254379
crossref_primary_10_1080_19312458_2024_2363776
crossref_primary_10_1177_08944393251366243
crossref_primary_10_1371_journal_pdig_0000631
crossref_primary_10_1017_rsm_2025_10031
crossref_primary_10_1177_00491241251338246
crossref_primary_10_1007_s40593_024_00414_0
crossref_primary_10_1111_pops_70004
crossref_primary_10_1016_j_jpi_2023_100338
crossref_primary_10_1371_journal_pone_0314136
crossref_primary_10_1162_tacl_a_00685
crossref_primary_10_1038_s42256_025_00986_z
crossref_primary_10_1093_nsr_nwaf169
crossref_primary_10_2139_ssrn_5342180
crossref_primary_10_7717_peerj_cs_2358
crossref_primary_10_3390_electronics13081409
crossref_primary_10_1007_s11747_025_01097_2
crossref_primary_10_1007_s41109_024_00658_8
crossref_primary_10_2196_50150
crossref_primary_10_1002_jcpy_1380
crossref_primary_10_1109_ACCESS_2024_3402809
crossref_primary_10_3390_electronics13081525
crossref_primary_10_1111_tgis_70111
crossref_primary_10_1007_s12599_023_00834_7
crossref_primary_10_1080_17489725_2025_2501632
crossref_primary_10_1007_s10207_024_00860_w
crossref_primary_10_1038_s41467_025_58344_x
crossref_primary_10_1162_opmi_a_00144
crossref_primary_10_1080_1369118X_2025_2531130
crossref_primary_10_1007_s11263_025_02392_9
crossref_primary_10_1109_ACCESS_2025_3525493
crossref_primary_10_3758_s13428_024_02381_9
crossref_primary_10_1038_s41598_024_53124_x
crossref_primary_10_1001_jamanetworkopen_2024_28276
crossref_primary_10_1007_s11192_025_05355_6
crossref_primary_10_3389_frai_2024_1365508
crossref_primary_10_1093_jamia_ocae197
crossref_primary_10_1080_00036846_2025_2532887
crossref_primary_10_1111_polp_12555
crossref_primary_10_1177_00491241251326819
crossref_primary_10_1007_s41060_024_00567_0
crossref_primary_10_1080_21645515_2024_2381925
crossref_primary_10_1108_EJIM_02_2024_0129
crossref_primary_10_1177_08944393251334977
crossref_primary_10_1002_ev_20557
crossref_primary_10_1073_pnas_2305016120
crossref_primary_10_1177_00491241251325243
crossref_primary_10_1007_s00146_025_02487_4
crossref_primary_10_1140_epjds_s13688_024_00456_3
crossref_primary_10_1057_s41599_025_05273_1
crossref_primary_10_1016_j_techfore_2025_124328
crossref_primary_10_20517_cf_2025_12
crossref_primary_10_1007_s42001_024_00272_9
crossref_primary_10_3390_app15042106
crossref_primary_10_1109_ACCESS_2024_3464374
crossref_primary_10_1109_MCI_2024_3504833
crossref_primary_10_3389_frsps_2024_1392128
crossref_primary_10_3390_app15148059
crossref_primary_10_2196_60678
crossref_primary_10_1177_08944393251349541
crossref_primary_10_1007_s40593_024_00431_z
crossref_primary_10_1007_s42001_025_00388_6
crossref_primary_10_1007_s40820_024_01423_3
crossref_primary_10_1007_s10639_025_13633_2
crossref_primary_10_1057_s41599_025_04503_w
crossref_primary_10_3390_info16050392
crossref_primary_10_1145_3676507
crossref_primary_10_1109_ACCESS_2023_3334260
crossref_primary_10_1080_10584609_2024_2364072
crossref_primary_10_3389_fpubh_2024_1352979
crossref_primary_10_1007_s10115_024_02321_1
crossref_primary_10_3389_fmed_2023_1259640
crossref_primary_10_1017_pan_2025_7
crossref_primary_10_1016_j_jss_2025_112533
crossref_primary_10_1177_16094069241231168
crossref_primary_10_1016_j_eswa_2025_126531
crossref_primary_10_1080_19312458_2025_2482538
crossref_primary_10_1162_coli_a_00561
crossref_primary_10_1109_ACCESS_2024_3354705
crossref_primary_10_1016_j_enpol_2025_114769
crossref_primary_10_1093_pnasnexus_pgae397
crossref_primary_10_1111_ecpo_12342
crossref_primary_10_1186_s12909_024_05239_y
crossref_primary_10_1007_s44212_025_00074_y
crossref_primary_10_1111_deci_12655
crossref_primary_10_1177_18344909251355673
crossref_primary_10_1038_s44387_025_00019_5
crossref_primary_10_1073_pnas_2314021121
crossref_primary_10_1145_3736402
crossref_primary_10_1007_s11432_023_3911_2
crossref_primary_10_1017_S0007123423000571
crossref_primary_10_1177_00491241241268453
crossref_primary_10_3758_s13428_024_02455_8
crossref_primary_10_1038_s44159_024_00392_z
crossref_primary_10_1038_s42256_023_00783_6
crossref_primary_10_1007_s10462_025_11328_1
crossref_primary_10_1155_atr_4874071
crossref_primary_10_1007_s42001_024_00345_9
crossref_primary_10_2196_64447
crossref_primary_10_1017_pan_2024_23
crossref_primary_10_1093_pnasnexus_pgae165
crossref_primary_10_1080_10503307_2025_2539405
crossref_primary_10_1177_00491241251342008
crossref_primary_10_1016_j_aei_2025_103244
crossref_primary_10_1145_3710947
crossref_primary_10_3390_ijgi14090325
crossref_primary_10_3390_ai6040072
crossref_primary_10_1038_s41598_025_16650_w
crossref_primary_10_1080_17502977_2024_2324567
crossref_primary_10_1080_10447318_2024_2392964
crossref_primary_10_2139_ssrn_5352879
crossref_primary_10_1177_2755323X251366728
crossref_primary_10_1098_rsos_250128
crossref_primary_10_1038_s41562_024_01938_0
crossref_primary_10_1038_s44387_025_00008_8
crossref_primary_10_1177_23780231241259651
crossref_primary_10_1017_pan_2024_31
crossref_primary_10_1097_JS9_0000000000001583
crossref_primary_10_1080_00380237_2025_2516814
crossref_primary_10_1016_j_envint_2025_109341
crossref_primary_10_1093_applin_amae071
crossref_primary_10_1177_13563890251330911
crossref_primary_10_1016_j_jad_2024_08_013
crossref_primary_10_1177_20563051251350975
crossref_primary_10_1080_15309576_2025_2539784
crossref_primary_10_1109_ACCESS_2024_3437197
crossref_primary_10_1145_3695993
crossref_primary_10_1080_19312458_2025_2549707
crossref_primary_10_1017_pan_2024_29
crossref_primary_10_3390_electronics14142800
crossref_primary_10_1016_j_ins_2024_121735
crossref_primary_10_1140_epjds_s13688_025_00548_8
crossref_primary_10_1026_0033_3042_a000699
crossref_primary_10_1002_advs_202508623
crossref_primary_10_1111_acem_14995
crossref_primary_10_1016_j_procs_2025_01_286
crossref_primary_10_1016_j_eswa_2024_125509
crossref_primary_10_1186_s40537_025_01236_0
crossref_primary_10_2196_51635
crossref_primary_10_3390_info16090812
crossref_primary_10_1177_00222429251314385
crossref_primary_10_1177_08944393241286471
crossref_primary_10_1111_lsq_12481
crossref_primary_10_1016_j_jmsy_2024_05_020
crossref_primary_10_1177_08944393251344865
crossref_primary_10_1177_00491241251339188
crossref_primary_10_1177_08944393251326175
crossref_primary_10_1080_19312458_2025_2500329
crossref_primary_10_1145_3649506
crossref_primary_10_1007_s10584_025_03959_8
crossref_primary_10_1080_13678868_2024_2336866
crossref_primary_10_1177_00491241251340608
crossref_primary_10_3390_make7030068
crossref_primary_10_1016_j_cmpb_2025_109070
crossref_primary_10_1016_j_rcim_2025_103076
crossref_primary_10_1002_csr_70148
crossref_primary_10_1098_rsos_241692
crossref_primary_10_3758_s13428_024_02337_z
crossref_primary_10_1007_s11192_025_05385_0
crossref_primary_10_1177_14614448241279034
crossref_primary_10_1038_s41698_025_01103_4
crossref_primary_10_1073_pnas_2309350120
crossref_primary_10_1088_3033_4942_addd43
crossref_primary_10_1109_ACCESS_2025_3554586
crossref_primary_10_1002_jcpy_1453
crossref_primary_10_1371_journal_pone_0331566
crossref_primary_10_1109_JPROC_2025_3572744
crossref_primary_10_1017_S0047279424000382
crossref_primary_10_1111_tgis_13282
crossref_primary_10_1017_psrm_2025_10010
crossref_primary_10_1017_S1537592724001907
crossref_primary_10_1109_TVCG_2024_3456373
crossref_primary_10_1016_j_jbi_2024_104724
crossref_primary_10_1109_ACCESS_2024_3349425
crossref_primary_10_32604_cmes_2024_052256
crossref_primary_10_1177_10422587241254069
crossref_primary_10_1111_jopy_13006
crossref_primary_10_1016_j_poetic_2024_101966
crossref_primary_10_1017_pan_2024_5
crossref_primary_10_1088_2631_8695_ad299e
crossref_primary_10_2196_65454
crossref_primary_10_1080_10810730_2024_2411320
crossref_primary_10_1016_j_esp_2024_11_003
crossref_primary_10_1038_s44284_025_00305_y
crossref_primary_10_1080_19312458_2025_2551693
crossref_primary_10_1109_TSE_2024_3356819
crossref_primary_10_1057_s41599_025_04877_x
crossref_primary_10_1016_j_knosys_2025_113386
crossref_primary_10_2139_ssrn_5357471
crossref_primary_10_1177_00104140241271297
crossref_primary_10_1007_s11558_024_09577_w
crossref_primary_10_1017_psrm_2024_35
crossref_primary_10_1093_pnasnexus_pgaf069
crossref_primary_10_1161_CIRCEP_124_013023
crossref_primary_10_3389_frsps_2025_1460277
crossref_primary_10_1007_s10664_024_10474_4
crossref_primary_10_1002_widm_1570
crossref_primary_10_3390_bdcc8040041
crossref_primary_10_1109_ACCESS_2024_3403426
crossref_primary_10_2139_ssrn_5317057
crossref_primary_10_1002_jrsm_1749
crossref_primary_10_1126_science_adq1814
crossref_primary_10_1057_s41599_024_02922_9
crossref_primary_10_1016_j_elerap_2025_101497
crossref_primary_10_1002_advs_202405789
crossref_primary_10_1080_08824096_2024_2410263
crossref_primary_10_1080_0144929X_2025_2551570
crossref_primary_10_3389_fcomm_2025_1610404
crossref_primary_10_1162_dint_a_00255
crossref_primary_10_1007_s11356_025_36566_2
crossref_primary_10_1109_TLT_2025_3570775
crossref_primary_10_1111_joms_13122
crossref_primary_10_1038_s43588_023_00585_1
crossref_primary_10_1177_00491241251336794
crossref_primary_10_1016_j_accinf_2024_100715
crossref_primary_10_3390_math13193062
crossref_primary_10_1016_j_osnem_2025_100319
crossref_primary_10_1038_s41598_025_98385_2
crossref_primary_10_3389_frai_2025_1602984
crossref_primary_10_1007_s10479_025_06713_6
crossref_primary_10_1162_qss_a_00360
crossref_primary_10_1016_j_jss_2025_112580
crossref_primary_10_1371_journal_pone_0313932
crossref_primary_10_1007_s42001_025_00362_2
crossref_primary_10_1016_j_landusepol_2025_107716
crossref_primary_10_1017_dap_2024_44
crossref_primary_10_1016_j_iswa_2023_200308
crossref_primary_10_1111_jcal_70089
crossref_primary_10_1057_s41304_024_00488_3
crossref_primary_10_1109_ACCESS_2025_3532995
crossref_primary_10_1177_00491241251333372
crossref_primary_10_3390_su16031166
crossref_primary_10_3233_WEB_230363
crossref_primary_10_1177_20531680241311510
crossref_primary_10_1109_JIOT_2024_3396282
crossref_primary_10_1057_s41599_025_05656_4
crossref_primary_10_1016_j_patter_2025_101341
crossref_primary_10_1109_ACCESS_2025_3540388
crossref_primary_10_1371_journal_pone_0302380
crossref_primary_10_1136_bmjopen_2023_082344
crossref_primary_10_1073_pnas_2401227121
crossref_primary_10_1371_journal_pone_0306621
crossref_primary_10_1017_S1049096525101248
crossref_primary_10_1080_19312458_2024_2383453
crossref_primary_10_1177_20531680251351910
crossref_primary_10_1007_s10579_025_09831_6
crossref_primary_10_1186_s12909_025_06881_w
crossref_primary_10_1145_3703454
crossref_primary_10_1038_s43588_025_00843_4
crossref_primary_10_1007_s42001_024_00338_8
crossref_primary_10_3390_app132011171
crossref_primary_10_1007_s11943_023_00332_y
ContentType Journal Article
DBID NPM
7X8
DOI 10.1073/pnas.2305016120
DatabaseName PubMed
MEDLINE - Academic
DatabaseTitle PubMed
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
PubMed
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Sciences (General)
EISSN 1091-6490
ExternalDocumentID 37463210
Genre Journal Article
GrantInformation_xml – fundername: EC | European Research Council (ERC)
  grantid: 883121
GroupedDBID ---
-DZ
-~X
.55
0R~
123
29P
2FS
2WC
4.4
53G
5RE
5VS
85S
AACGO
AAFWJ
AANCE
ABOCM
ABPLY
ABPPZ
ABTLG
ABZEH
ACGOD
ACIWK
ACNCT
ACPRK
AENEX
AFFNX
AFOSN
AFRAH
ALMA_UNASSIGNED_HOLDINGS
BKOMP
CS3
D0L
DIK
DU5
E3Z
EBS
F5P
FRP
GX1
H13
HH5
HYE
JLS
JSG
KQ8
L7B
LU7
N9A
NPM
N~3
O9-
OK1
PNE
PQQKQ
R.V
RHI
RNA
RNS
RPM
RXW
SJN
TAE
TN5
UKR
W8F
WH7
WOQ
WOW
X7M
XSW
Y6R
YBH
YKV
YSK
ZCA
~02
~KM
7X8
JENOY
ID FETCH-LOGICAL-c524t-1343d242301ef128068e62f36fef24d26aa00b47e9004c07bf288c425550b1db2
IEDL.DBID 7X8
ISICitedReferencesCount 391
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001083464000007&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1091-6490
IngestDate Wed Oct 01 13:59:50 EDT 2025
Thu Apr 03 06:58:24 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 30
Keywords human annotations
large language models
text classification
ChatGPT
text as data
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c524t-1343d242301ef128068e62f36fef24d26aa00b47e9004c07bf288c425550b1db2
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ORCID 0000-0001-6696-6471
0000-0002-5592-9648
0000-0002-0635-3048
OpenAccessLink https://www.pnas.org/doi/10.1073/pnas.2305016120
PMID 37463210
PQID 2839741439
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2839741439
pubmed_primary_37463210
PublicationCentury 2000
PublicationDate 2023-07-25
PublicationDateYYYYMMDD 2023-07-25
PublicationDate_xml – month: 07
  year: 2023
  text: 2023-07-25
  day: 25
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle Proceedings of the National Academy of Sciences - PNAS
PublicationTitleAlternate Proc Natl Acad Sci U S A
PublicationYear 2023
SSID ssj0009580
Score 2.7630637
Snippet Many NLP applications require manual text annotations for a variety of tasks, notably to train classifiers or evaluate the performance of unsupervised models....
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
StartPage e2305016120
Title ChatGPT outperforms crowd workers for text-annotation tasks
URI https://www.ncbi.nlm.nih.gov/pubmed/37463210
https://www.proquest.com/docview/2839741439
Volume 120
WOSCitedRecordID wos001083464000007&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LS8QwEA7qevCirs_1RQQPeijbJmnT4EFkcfXisocV9lbSPFCEtpqu_n0nbRe9CIKXHgqBMsk383VmMh9CF0YSEdlYA75DHTDCbCA55UFujdAhz2mi80Zsgk8m6Xwupl3CzXVtlUuf2DhqXSqfIx9CGATqC9Fd3FRvgVeN8tXVTkJjFfUoUBkPTD5PfwzdTdtpBCIKEibC5WgfTodVIZ1vgo495SHh7_yyiTPjrf9-4Tba7Bgmvm2PRB-tmGIH9TsMO3zZDZq-2kXXo2dZ309nuFzUVXuBwGFwzJ8a-3YtIIYY3mHfGxLIoijbqj2upXt1e-hpfDcbPQSdmEKgYsK85Dyj2pOnMDI28vXU1CTE0sQaS5gmiZRhmDNuBMBGwS5ZkqYKEA2_MHmkc7KP1oqyMIcIKx3DJjJNwTUySnNYmeo0FEoyoaNEDdD50kAZHFZfgZCFKRcu-zbRAB20Vs6qdqpGRjlL_IWioz-sPkYbXvbd51hJfIJ6FqBqTtG6-qhf3PtZcwrgOZk-fgFS8rwP
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=ChatGPT+outperforms+crowd+workers+for+text-annotation+tasks&rft.jtitle=Proceedings+of+the+National+Academy+of+Sciences+-+PNAS&rft.au=Gilardi%2C+Fabrizio&rft.au=Alizadeh%2C+Meysam&rft.au=Kubli%2C+Ma%C3%ABl&rft.date=2023-07-25&rft.eissn=1091-6490&rft.volume=120&rft.issue=30&rft.spage=e2305016120&rft_id=info:doi/10.1073%2Fpnas.2305016120&rft_id=info%3Apmid%2F37463210&rft_id=info%3Apmid%2F37463210&rft.externalDocID=37463210
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1091-6490&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1091-6490&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1091-6490&client=summon