JMIR JMIR AI JMIR AI 2817-1705 JMIR Publications Toronto, Canada v3i1e53656 38875596 10.2196/53656 Research Letter Research Letter What Is the Performance of ChatGPT in Determining the Gender of Individuals Based on Their First and Last Names? El Emam Khaled Malin Bradley Teel Zoë Abbie Shamsi Amrollah Zhu Lingxuan Sebo Paul MSc, MD 1
University Institute for Primary Care University of Geneva Rue Michel-Servet 1 Geneva, 1211 Switzerland 41 223794390 paulsebo@hotmail.com
https://orcid.org/0000-0001-7616-0017 University Institute for Primary Care University of Geneva Geneva Switzerland Corresponding Author: Paul Sebo paulsebo@hotmail.com 2024 13 3 2024 3 e53656 14 10 2023 21 11 2023 26 11 2023 2 3 2024 ©Paul Sebo. Originally published in JMIR AI (https://ai.jmir.org), 13.03.2024. 2024

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR AI, is properly cited. The complete bibliographic information, a link to the original publication on https://www.ai.jmir.org/, as well as this copyright and license information must be included.

 accuracy artificial intelligence AI ChatGPT gender gender detection tool misclassification name performance gender detection gender detection tools inequalities language model NamSor Gender API Switzerland physicians gender bias disparities gender disparities gender gap Introduction

Accurate determination of gender from names is vital for addressing gender-related disparities in medicine and promoting inclusivity. Gender detection tools (GDTs) offer efficient solutions, enabling large-scale demographic analysis [1-3] to improve data quality and inform targeted interventions. Indeed, they can process thousands of names simultaneously, saving time and resources. However, most of them charge for more than a certain number of requests per month. We recently compared the performance of 4 GDTs and showed that Gender API (Gender-API.com) and NamSor (NamSor Applied Onomastics) were accurate (misclassifications=1.5% and 2.0%, respectively; nonclassifications=0.3% and 0%, respectively) [4].

ChatGPT is a language model developed by OpenAI that is capable of generating human-like text and engaging in natural language conversations [5]. In medicine, ChatGPT can be employed for various purposes, such as answering patient queries and providing information on medical topics, making it a valuable resource for health care professionals and researchers seeking quick access to medical information and support in their work [6,7].

Given the increasing usefulness of GDTs in research, particularly for evaluating gender disparities in medicine, we assessed whether the performance of ChatGPT as a free GDT (version GPT-3.5) could approach that of Gender API and NamSor. We also compared ChatGPT-3.5 with the more advanced GPT-4 version. We hypothesized that ChatGPT, a versatile language model not specifically trained for gender analysis, could achieve gender detection performance comparable to specialized tools and that ChatGPT-4 would perform no better than ChatGPT-3.5.

 Methods Database Selection and Data Collection

The methods used in this study are the same as those used in our primary study, which compared the performance of 4 GDTs [4]. We used a database of 6131 physicians practicing in Switzerland, a multilingual and multicultural country with 36% of physicians of foreign origin [4]. The sample consisted of 3085 women (50.3%) and 3046 men (49.7%), with gender determined by self-identification. We used nationalize.io to determine the origin of physicians’ names (Table 1). A total of 88% of names were from French-, English-, Spanish-, Italian-, German-, or Portuguese-speaking countries or from another European country.

We asked ChatGPT-3.5 to determine the gender of 500 physicians at a time, after copying and pasting these lists of first and last names from the database. We ran the analysis twice and also examined ChatGPT-4 to check the “stability” of the responses [8]. The data were collected between September and November 2023.

We constructed a confusion matrix (Table 2): ff and mm correspond to correct classifications, mf and fm to misclassifications, and fu and mu to nonclassifications (ie, gender impossible to determine).

As in other studies [4,9], we calculated 4 performance metrics, namely “errorCoded” (the proportion of misclassifications and nonclassifications), “errorCodedWithoutNA” (the proportion of misclassifications), “naCoded” (the proportion of nonclassifications), and “errorGenderBias” (the direction of bias in gender determination). We used Cohen κ to assess interrater agreement.

Estimated origin of physicians’ names (N=6131 physicians).

Origin Counta, n (%)
French-speaking country 1679 (32.2)
English-speaking country 751 (14.4)
Spanish-speaking country 404 (7.7)
Asian countryb 344 (6.6)
Eastern European country 324 (6.2)
Italian-speaking country 288 (5.5)
Western European countryb 272 (5.2)
Arabic-speaking country 259 (5.0)
German-speaking country 259 (5.0)
Northern European countryb 220 (4.2)
Southern European countryb 217 (4.2)
Portuguese-speaking country 198 (3.8)

aThe total number of physicians does not add to 6131 because of missing values (no assignments for 916 physicians).

bIf not already classified in another group (eg, in the Arabic-speaking country group for some Asian countries).

Confusion matrix showing the 6 possible classification outcomes.

Female (predicted) Male (predicted) Unknown (predicted)
Female (actual) ff fm fu
Male (actual) mf mm mu
Ethical Considerations

Since this study did not involve the collection of personal health–related data, it did not require ethical review per current Swiss law.

 Results

Performance metrics showed high accuracy for ChatGPT-3.5 and ChatGPT-4 in both the first and second rounds (Table 3). The number of misclassifications was low (proportion≤1.5%) and there were no “nonclassifications.” As shown in Table 3, interrater agreement between the first and second rounds (for ChatGPT-3.5 and ChatGPT-4) and between ChatGPT-3.5 and ChatGPT-4 (for the first round) was “almost perfect” (κ>0.97, all Ps<.001).

Confusion matrix and performance metrics for ChatGPT-3.5 and ChatGPT-4 (N=6131 physicians).

Classified as women, n (%) Classified as men, n (%) Unclassified, n (%) Interrater agreementa
Cohen κ (95% CI) P value
ChatGPT-3.5 0.9817 (0.9770-0.9865)b <.001
First roundc
Female physicians (n=3085) 3028 (98.2) 57 (1.8) 0 (0)
Male physicians (n=3046) 18 (0.6) 3028 (99.4) 0 (0)
Second roundd
Female physicians (n=3085) 3030 (98.2) 55 (1.8) 0 (0)
Male physicians (n=3046) 28 (0.9) 3018 (99.1) 0 (0)
ChatGPT-4 0.9958 (0.9935-0.9981)b <.001
First rounde
Female physicians (n=3085) 3020 (97.9) 65 (2.1) 0 (0)
Male physicians (n=3046) 27 (0.9) 3019 (99.1) 0 (0)
Second roundf
Female physicians (n=3085) 3020 (97.9) 65 (2.1) 0 (0)
Male physicians (n=3046) 26 (0.9) 3020 (99.1) 0 (0)

aInterrater agreement between ChatGPT-3.5 and ChatGPT-4 (for the first round): Cohen κ=0.9768, 95% CI 0.9715-0.9822, P<.001.

bInterrater agreement between the first and second rounds for each version.

cPerformance metrics: errorCoded=0.01223, errorCodedWithoutNA=0.01223, naCoded=0, and errorGenderBias=–0.00636.

dPerformance metrics: errorCoded=0.01354, errorCodedWithoutNA=0.01354, naCoded=0, and errorGenderBias=–0.00440.

ePerformance metrics: errorCoded=0.01501, errorCodedWithoutNA=0.01501, naCoded=0, and errorGenderBias=–0.00620.

fPerformance metrics: errorCoded=0.01484, errorCodedWithoutNA=0.01484, naCoded=0, and errorGenderBias=–0.00636.

 Discussion

We used ChatGPT to determine the gender of 6131 physicians practicing in Switzerland and found that the proportion of misclassifications was ≤1.5% for both versions. There were no nonclassifications and gender bias was negligible. Interrater agreement between ChatGPT-3.5 and ChatGPT-4 was “almost perfect.”

These results are relatively similar to those found in our primary study for Gender API and NamSor (errorCoded=0.0181 and 0.0202, errorCodedWithoutNA=0.0147 and 0.0202, naCoded=0.0034 and 0, errorGenderBias=–0.0072 and 0.0026) [4]. They are slightly better than those of another study published in 2018, which compared 5 GDTs, including Gender API and NamSor [9]. These results suggest that ChatGPT can accurately determine the gender of individuals using their first and last names. The disadvantage of ChatGPT compared to Gender API and NamSor is that the database cannot be uploaded directly into ChatGPT (eg, as an Excel or CSV file).

Both ChatGPT-3.5 and ChatGPT-4 exhibit high accuracy in gender detection, with no significant superiority observed in ChatGPT-4 over ChatGPT-3.5. This underscores the robustness of ChatGPT in gender prediction across different versions. Our short study has 2 main limitations. Given the estimated origin of physicians’ names, the results of the study can probably be generalized to most Western countries but not necessarily to Asian or Middle Eastern countries. GDTs are often less accurate with names from these countries [9,10]. In addition, GDTs oversimplify the concept of gender by dichotomizing individuals into male or female.

 Abbreviations GDT

gender detection tool

 Data Availability

The data associated with this article are available in the Open Science Framework [11].

None declared.

 Cevik M Haque S Manne-Goehler J Kuppalli Krutika Sax Paul E Majumder Maimuna S Orkin Chloe Gender disparities in coronavirus disease 2019 clinical trial leadership Clin Microbiol Infect 2021 07 27 7 1007 1010 10.1016/j.cmi.2020.12.025 33418021 S1198-743X(20)30785-0 PMC7785275 Sebo P Clair C Gender gap in authorship: a study of 44,000 articles published in 100 high-impact general medical journals Eur J Intern Med 2022 03 97 103 105 10.1016/j.ejim.2021.09.013 34598855 S0953-6205(21)00313-7 Gottlieb M Krzyzaniak SM Mannix A Parsons M Mody S Kalantari A Ashraf H Chan TM Sex distribution of editorial board members among emergency medicine journals Ann Emerg Med 2021 01 77 1 117 123 10.1016/j.annemergmed.2020.03.027 32376090 S0196-0644(20)30217-1 Sebo P Performance of gender detection tools: a comparative study of name-to-gender inference services J Med Libr Assoc 2021 07 01 109 3 414 421 10.5195/jmla.2021.1185 34629970 jmla.2021.1185 PMC8485937 Rao A Pang M Kim J Kamineni M Lie W Prasad AK Landman A Dreyer K Succi MD Assessing the utility of ChatGPT throughout the entire clinical workflow: development and usability study J Med Internet Res 2023 08 22 25 e48659 10.2196/48659 37606976 v25i1e48659 PMC10481210 Dave T Athaluri SA Singh S ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations Front Artif Intell 2023 5 4 6 1169595 10.3389/frai.2023.1169595 37215063 PMC10192861 Sallam M ChatGPT utility in healthcare education, research, and practice: systematic review on the promising perspectives and valid concerns Healthcare (Basel) 2023 03 19 11 6 887 10.3390/healthcare11060887 36981544 healthcare11060887 PMC10048148 Zhu L Mou W Yang T Chen R ChatGPT can pass the AHA exams: open-ended questions outperform multiple-choice format Resuscitation 2023 07 188 109783 10.1016/j.resuscitation.2023.109783 37349064 S0300-9572(23)00096-5 Santamaría Lucía Mihaljević Helena Comparison and benchmark of name-to-gender inference services PeerJ Comput Sci 2018 4 e156 10.7717/peerj-cs.156 33816809 cs-156 PMC7924484 Sebo P How accurate are gender detection tools in predicting the gender for Chinese names? A study with 20,000 given names in Pinyin format J Med Libr Assoc 2022 04 01 110 2 205 211 10.5195/jmla.2022.1289 35440899 jmla.2022.1289 PMC9014919 What is the performance of ChatGPT in determining the gender of individuals based on their first and last names? Open Science Framework 2023 09 27 2024-03-08 https://osf.io/6nzd4/