The DOI matches its Crossref record; 53 of the 66 references it cites were found in a registry or an index.
Deep neural networks are more accurate than humans at detecting sexual orientation from facial images.
Open accessPublished version of a preprint
Abstract
We show that faces contain much more information about sexual orientation than can be perceived or interpreted by the human brain. We used deep neural networks to extract features from 35,326 facial images. These features were entered into a logistic regression aimed at classifying sexual orientation. Given a single facial image, a classifier could correctly distinguish between gay and heterosexual men in 81% of cases, and in 71% of cases for women. Human judges achieved much lower accuracy: 61% for men and 54% for women. The accuracy of the algorithm increased to 91% and 83%, respectively, given five facial images per person. Facial features employed by the classifier included both fixed (e.g., nose shape) and transient facial features (e.g., grooming style). Consistent with the prenatal hormone theory of sexual orientation, gay men and women tended to have gender-atypical facial morphology, expression, and grooming styles. Prediction models aimed at gender alone allowed for detecting gay males with 57% accuracy and gay females with 58% accuracy. Those findings advance our understanding of the origins of sexual orientation and the limits of human perception. Additionally, given that companies and governments are increasingly using computer vision algorithms to detect people's intimate traits, our findings expose a threat to the privacy and safety of gay men and women. (PsycINFO Database Record
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“For example, a recent paper used publicly available images of faces uploaded to a dating website as a way to train AI software to detect someone’s sexuality based on a small number of photos (Wang and Kosinski 2018). While the study received ethics committee approval, it raises further questions around consent, since it is implausible that the users of the dating website could or necessarily would”
From How to Design AI for Social Good: Seven Essential Factors (2020) · cites this paper as Wang and Kosinski 2018
In the citing paper
“For example, a study found that computer vision algorithms were used to ascertain sexual orientation of people by scrutinizing thousands of facial images taken from public profiles on a dating website (Wang and Kosinski, 2017).”
From SHIFTing artificial intelligence to be responsible in healthcare: a systematic review (2022) · cites this paper as Wang and Kosinski, 2017
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Dermatologist-level classification of skin cancer with deep neural networks
Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017) · Nature · doi:10.1038/nature21056 · Cited by 14,362
Parkhi, O. M., Vedaldi, A., & Zisserman, A. (2015) · Procedings of the British Machine Vision Conference 2015 · doi:10.5244/c.29.41 · Cited by 5,479
Private traits and attributes are predictable from digital records of human behavior
Kosinski, M., Stillwell, D., & Graepel, T. (2013) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1218772110 · Cited by 2,829
Self-citationThe Evolutionary Psychology of Facial Beauty
Rhodes, G. (2005) · Annual Review of Psychology · doi:10.1146/annurev.psych.57.102904.190208 · Cited by 1,821
Inferences of competence from faces predict election outcomes
Todorov, A., Mandisodza, A. N., Goren, A., & Hall, C. C. (2005) · Science · doi:10.1126/science.1110589 · Cited by 1,508
Social Cognition: Thinking Categorically about Others
Macrae, C. N., & Bodenhausen, G. V. (2000) · Annual Review of Psychology · doi:10.1146/annurev.psych.51.1.93 · Cited by 1,432
The evolution of human skin coloration
Jablonski, N. G., & Chaplin, G. (2000) · Journal of Human Evolution · doi:10.1006/jhev.2000.0403 · Cited by 1,212
Kosiński, M., Matz, S., Gosling, S. D., Popov, V., & Stillwell, D. J. (2015) · American Psychologist · doi:10.1037/a0039210 · Cited by 1,031
Self-citation
Cited by
Sorted by citations
Items 201–300 of 778
The Profiling Potential of Computer Vision and the Challenge of Computational Empiricism
Goldenfein, J. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287568 · Cited by 21
Giantini, G. (2023) · AI and Ethics · doi:10.1007/s43681-023-00311-7 · Cited by 13
“Recently, Deep Neural Networks, another AI subcategory, have been trained to perform this function, as in a study based on the analysis of a set of 35,326 facial images obtained from a dating website [51].”
BackgroundMethodKrupiy, T., & Scheinin, M. (2023) · Human Rights Law Review · doi:10.1093/hrlr/ngad019 · Cited by 18
Ullstein, C., Engelmann, S., Papakyriakopoulos, O., Ikkatai, Y., Arnez-Jordan, N. P., Caleno, R., Mboya, B., Higuma, S., Hartwig, T., Yokoyama, H., et al. (2024) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3630106.3659038 · Cited by 14
“…measures [11], research projects in facial analysis have claimed their models can reliably infer people’s ethnicity [67], sexual orientation [58, 101], political ideology and orientation [53, 77, 104], emotion expression and intensity [8, 24], or personality traits [5, 85] based on facial…”
BackgroundVoice and Prejudice: The Social Costs of Auditory Gaydar
Fasoli, F., & Maass, A. (2018) · Atlantic Journal of Communication · doi:10.1080/15456870.2018.1432617 · Cited by 13
Social media and its intersections with free speech, freedom of information and privacy. An analysis
Segado‐Boj, F., & Díaz-Campo, J. (2020) · Revista ICONO14 · doi:10.7195/ri14.v18i1.1379 · Cited by 13
“Social media companies should clearly comply with the legal requirements in each national setting to forbid unlawful material to spread through their users profiles.”
BackgroundFacing AI: conceptualizing ‘fAIce communication’ as the modus operandi of facial recognition systems
Bucher, T. (2022) · Media Culture & Society · doi:10.1177/01634437211036975 · Cited by 13
“…from dating sites and training a sample of over 35,000 facial images of self-identified gay and straight individuals, Wang and Kosinski claimed the algorithm could correctly distinguish between gay and straight people 91% of the time for men and 83% of the time for women (Wang and Kosinski, 2018).”
BackgroundArtificial intimacy: ethical issues of AI romance
Shank, D. B., Koike, M., & Loughnan, S. (2025) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2025.02.007 · Cited by 19
Regulating around freedom in the “forum internum”
Alegre, S. (2020) · ERA Forum · doi:10.1007/s12027-020-00633-7 · Cited by 12
Item Response Theory: A Statistical Framework for Educational and Psychological Measurement
Chen, Y., Li, X., Liu, J., & Ying, Z. (2025) · Statistical Science · doi:10.1214/23-sts896 · Cited by 30
“…researchers tend to study psychological constructs using not only item response data but also a much wider range of daily behavioral data including those from social media, facial expression, and writing and speech (e.g., Inkster et al., 2016; Nave et al., 2018; Wang and Kosinski, 2018).”
Smart Festivals? Security and Freedom for Well-Being in Urban Smart Spaces
Crampton, J. W., Hoover, K. C., Smith, H., et al. (2019) · Annals of the American Association of Geographers · doi:10.1080/24694452.2019.1662765 · Cited by 12
“Third, leading-edge biometric technologies promise to infer not just identity (who someone is) but characteristics of people’s nature such as sexual orientation (Wang and Kosinski 2018) and criminality (Wu and Zhang 2017).”
BackgroundGoldenfein, J. (2019) · Cambridge University Press eBooks · doi:10.1017/9781108637657 · Cited by 14
Vista: Optimized System for Declarative Feature Transfer from Deep CNNs at Scale
Nakandala, S., & Kumar, A. (2020) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3318464.3389709 · Cited by 11
“Such feature transfer underpins recent breakthrough applications of CNNs in detecting cancer [34], diabetic retinopathy [63], facial analysis [66], and multimodal recommendation systems [53].”
BackgroundResultA Short Introduction to the Ethics of Artificial Intelligence
Liao, S. M. (2020) · Oxford University Press eBooks · doi:10.1093/oso/9780190905033.003.0001 · Cited by 13
Respect as a Lens for the Design of AI Systems
Seymour, W., Kleek, M. V., Binns, R., & Murray-Rust, D. (2022) · AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3514094.3534186 · Cited by 18
“sexual orientation [101] or political persuasion [63] raises many concerns [78]; simplifying the questions to ones rooted in respect for persons and recognition respect – “would the person in this image want to be classified this way”, “how does this classification affect the”
BackgroundPereira, K. J., David, V. F., Varella, M. A. C., & Valentová, J. V. (2020) · Evolution and Human Behavior · doi:10.1016/j.evolhumbehav.2020.05.003 · Cited by 12
Community attitudes on genetic research of gender identity, sexual orientation, and mental health
Thomas, T., Hofammann, D., McKenna, B. G., van der Miesen, A. I. R., Stokes, M. A., Daniolos, P. T., & Michaelson, J. J. (2020) · PLoS ONE · doi:10.1371/journal.pone.0235608 · Cited by 14
Modeling Ethics: Approaches to Data Creep in Higher Education
Whitman, M. (2021) · Science and Engineering Ethics · doi:10.1007/s11948-021-00346-1 · Cited by 15
“An example: in 2017, Stanford researchers used profile photos and data from a dating app to predict sexuality based on faces (see Wang & Kosinski, 2017).”
MethodPitfalls in Machine Learning Research: Reexamining the Development CyclePreprint
Biderman, S., & Scheirer, W. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.02832 · Cited by 19
“Recent examples of this include algorithms for determining criminal tendencies [53, 22] and detecting sexual orientation [51] from photos of faces, as well as commercial services that automatically determine suitable job candidates for hiring managers [12].”
Post-Privacy oder der Verlust der Informationskontrolle
Hagendorff, T. (2019) · J.B. Metzler eBooks · doi:10.1007/978-3-476-04860-8_6 · Cited by 11
Socialni odnosi, animizem, antropomorfizem in interakcija z UI
Babić, N. Č. (2022) · doi:10.18690/um.ff.11.2022.12 · Cited by 10
Quantifying the Dialect Gap and its Correlates Across Languages
Kantharuban, A., Vulić, I., & Korhonen, A. (2023) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2023.findings-emnlp.481 · Cited by 22
Gaming faces: diagnostic scanning in social media and the legacy of racist face analysis
Monteiro, S. (2022) · Information Communication & Society · doi:10.1080/1369118x.2021.2020867 · Cited by 11
“…the possibility of producing algorithms capable of detecting personality traits, sexual orientation, ethnicity, or predicting an individual ’ s propensity for criminality or violence based on face shape, facial features, or expression (Wang & Kosinski, 2018; Wang et al., 2018; Wu & Zhang, 2016).”
BackgroundMaliepaard, E., & Baumgartner, R. R. (2020) · Routledge eBooks · doi:10.4324/9780367809881 · Cited by 16
AI-Competent Individuals and Laypeople Tend to Oppose Facial Analysis AI
Ullstein, C., Engelmann, S., Papakyriakopoulos, O., Hohendanner, M., & Großklags, J. (2022) · ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO) · doi:10.1145/3551624.3555294 · Cited by 12
Deep Learning for Political Science
Chatsiou, K., & Mikhaylov, S. (2020) · SAGE Publications Ltd eBooks · doi:10.4135/9781526486387.n58 · Cited by 19
“n et al., 2016) provided evidence that a risk-assessment machine learning algorithm used by US courts wrongly flagged non-white defendants at almost twice the rate of white defendants. More recently, Wang and Kosinski (2018) showed how deep neural networks can outperform humans in detecting sexual orientation. Apart f…”
BackgroundJones, J. A. B., Nash, U. W., VIEILLEFONT, J., et al. (2019) · Scientific Data · doi:10.1038/s41597-019-0245-9 · Cited by 11
Wang, Y. (2023) · doi:10.36227/techrxiv.23968311 · Cited by 13
Self-citationEl big data en los procesos políticos: hacia una democracia de la vigilancia
García, C. S. (2023) · Revista de filosofía · doi:10.4067/s0718-43602023000100215 · Cited by 10
How AI hype impacts the LGBTQ+ community
McAra-Hunter, D. (2024) · AI and Ethics · doi:10.1007/s43681-024-00423-8 · Cited by 14
“One notable instance occurred in 2017 when a Stanford study claimed that facial images could be used by artificial intelligence to determine sexual orientation [134].”
BackgroundArtificial Intelligence and Human Rights
(2023) · Oxford University Press eBooks · doi:10.1093/law/9780192882486.001.0001 · Cited by 10
An AI Method to Score Celebrity Visual Potential
Feng, X., Zhang, S., Liu, X., Srinivasan, K., & Lamberton, C. P. (2021) · Journal of Marketing Research · doi:10.1177/00222437251323238 · Cited by 15
“…economic value (Grewal, Gupta, and Hamilton 2021; Li and Xie 2019; Zhang and Luo 2022; Zhang et al. 2021b; Zhou, Lu, and Ding 2020), among which the human face is a popular topic (Hartmann et al. 2021; Hu and Ma 2021; Kachur et al. 2020; Peng et al. 2022; Wang and Kosinski 2018; Zhou et al. 2021).”
The Ethics of Uncertainty for Data Subjects
Nickel, P. J. (2019) · Philosophical studies series · doi:10.1007/978-3-030-04363-6_4 · Cited by 12
Deep Neural Networks and Explainable Machine Learning
Maratea, A., & Ferone, A. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-12544-8_23 · Cited by 10
Affective computing in the context of music therapy: a systematic review
de Santana, M. A., de Lima, C. L., Torcate, A. S., Fonseca, F. S., & dos Santos, W. P. (2021) · Research Society and Development · doi:10.33448/rsd-v10i15.22844 · Cited by 11
“Another popular method for emotion recognition is through facial expressions (Y. Wang & Kosinski, 2018).”
MethodMadan, S., Savani, K., & Johar, G. V. (2022) · Journal of Personality and Social Psychology · doi:10.1037/pspa0000307 · Cited by 11
“…infer numerous aspects of people’s character just from their faces (i.e., facial profiling), including traits such as criminality (Wu & Zhang, 2016), sexual orientation (Wang & Kosinski, 2018), political orientation (Kosinski, 2021), and openness, agreeableness, and neuroticism (Suen et al., 2020).”
BackgroundAn Essay on Human Values in HCI
Pereira, R., Baranauskas, M. C. C., & Liu, K. (2018) · Journal on Interactive Systems · doi:10.5753/jis.2018.689 · Cited by 13
Psychological trait inferences from women’s clothing: human and machine prediction
Rosenbusch, H., Aghaei, M., Evans, A. M., & Zeelenberg, M. (2020) · Journal of Computational Social Science · doi:10.1007/s42001-020-00085-6 · Cited by 12
Schmid, I., Witkower, Z., Götz, F. M., & Stieger, S. (2022) · Scientific Reports · doi:10.1038/s41598-022-22709-9 · Cited by 11
Statistical Data Analysis in the Era of Big Data
Lengauer, T. (2020) · Chemie Ingenieur Technik · doi:10.1002/cite.202000024 · Cited by 11
“Or intimate aspects of the personality structure can be derived from facial images by means of data analysis [35].”
BackgroundHirst, G., Yeo, G., Celestine, N. A., Lin, S.-Y., & Richardson, A. (2020) · Australian Journal of Management · doi:10.1177/0312896220919468 · Cited by 11
“If sexual preferences can already be predicted by neural networks (Wang and Kosinski, 2018), genetic technologies can be used to anticipate behaviour, and workplaces can be constructed to control these.”
BackgroundImpact of Biases in Big DataPreprint
Glauner, P., Valtchev, P., & State, R. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.00897 · Cited by 22
“Second, these machine learning models may discriminate against subjects of under-represented categories [3, 4].”
BackgroundExamining the Presence of Gender Bias in Customer Reviews Using Word EmbeddingPreprint
Mishra, A., Mishra, H., & Rathee, S. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3327404 · Cited by 16
Building Human Rights by Design
Kunnathully, K., Jinesh, M., Satyanarayana, P., Esther, V., Pathak, G. P., Selvakumar, P., & Manjunath, T. C. (2025) · IGI Global eBooks · doi:10.4018/979-8-3373-3830-9.ch004 · Cited by 9
The hard proxy problem: proxies aren’t intentional; they’re intentional
Johnson, G. M. (2025) · Philosophical Studies · doi:10.1007/s11098-025-02333-9 · Cited by 11
“This criticism echoes a fundamental lesson argued for by leading philosophical figures in mind and language, such as Putnam ( 1975a,b), Kripke (1980), and Burge 51 Wang & Kosinski, 2018.”
An introductory guide for conducting psychological research with big data.
Vezzoli, M., & Zogmaister, C. (2023) · Psychological Methods · doi:10.1037/met0000513 · Cited by 9
Perceiving education from Facebook profile pictures
Reiss, M. V., & Tsvetkova, M. (2019) · New Media & Society · doi:10.1177/1461444819868678 · Cited by 8
“Furthermore, photographs convey enough information to help us, sometimes aided by machines, guess better than chance a person’s sexual orientation (Wang and Kosinski, 2018), political affiliation (Rule and Ambady, 2010), their criminality and corruptibility (Lin et al., 2018; Valla et al., 2011),…”
BackgroundInfluentialYu, S., & Carroll, F. (2022) · Springer eBooks · doi:10.1007/978-3-031-06709-9_9 · Cited by 11
Klumbytė, G., Lücking, P., & Draude, C. (2020) · Nordic Conference on Human-Computer Interaction · doi:10.1145/3419249.3420120 · Cited by 11
“…instance, what kind of algorithmic experience would be had by subjects of algorithms such as “AI gaydar” that claimed to be able to classify people as straight or gay based on their dating profile photos [64], or an algorithm that suggests inference of criminality from images of faces alone [65]?”
BackgroundResearch Ethics, Vulnerability, and Trust on the Internet
Tiidenberg, K. (2018) · Springer eBooks · doi:10.1007/978-94-024-1202-4_55-1 · Cited by 10
Assessing Gender Bias in Machine Translation -- A Case Study with Google Translate
Prates, M. O. R., Avelar, P. H. C., & Lamb, L. (2018) · arXiv · doi:10.48550/arxiv.1809.02208 · Cited by 409
“ectly revived heated debates about the controversial eld of physiognomy, with proposals of AI systems capable of identifying the sexual orientation of an individual through its facial characteristics [38]. Similar concerns are growing at an unprecedented rate in the media, with reports of Apple’s Iphone X face unlock…”
BackgroundAssessing the Big Five personality traits using real-life static facial imagesPreprint
Kachur, A., Осин, Е. Н., Davydov, D., Shutilov, K., & Novokshonov, A. (2020) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/3y98a · Cited by 9
Foucault’s error: The power of not knowing
Offer, K., Rahwan, Z., & Hertwig, R. (2024) · European Review of Social Psychology · doi:10.1080/10463283.2024.2383085 · Cited by 11
“For example, Wang and Kosinski (2018) developed an algorithm that uses dating website data to detect an individual’s sexual orientation based on five facial images per person.”
Ziegler, M. D., & Peikert, A. (2018) · Journal of Intelligence · doi:10.3390/jintelligence6030041 · Cited by 8
AEGAN: Generating imperceptible face synthesis via autoencoder‐based generative adversarial network
Che, A., Yang, J., Guo, C., et al. (2023) · Computer Animation and Virtual Worlds · doi:10.1002/cav.2160 · Cited by 8
“6,7 Consequently, powerful FR systems have been exploited to identify the faces of victims, thereby illegally obtaining victims’ privacy information.”
BackgroundThe Group Right to Mutual Privacy
Puri, A. (2023) · Digital Society · doi:10.1007/s44206-023-00051-5 · Cited by 9
Jantunen, J. H. (2018) · AFinLA-e Soveltavan kielitieteen tutkimuksia · doi:10.30660/afinla.69259 · Cited by 8
Positive association between facial and vocal femininity/masculinity in women but not in men
Pereira, K. J., Varella, M. A. C., Kleisner, K., et al. (2019) · Behavioural Processes · doi:10.1016/j.beproc.2019.04.010 · Cited by 8
Image Perturbation-Based Deep Learning for Face Recognition Utilizing Discrete Cosine Transform
Park, J., & Kim, K. (2021) · Electronics · doi:10.3390/electronics11010025 · Cited by 9
“In addition, information more than just the person’s identity can be inferred from the feature representations of face recognition [6,7].”
BackgroundAn Ethical Highlighter for People-Centric Dataset CreationPreprint
Hanley, M., Khandelwal, A., Averbuch‐Elor, H., Snavely, N., & Nissenbaum, H. F. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.13583 · Cited by 18
“For example, a detractor may assert that the purpose for creating a face dataset being to tell apart gay individuals does not meet a moral threshold [20].”
BackgroundKaufman, D., Junquilho, T. A., & Reis, P. (2023) · Revista de Direitos e Garantias Fundamentais · doi:10.18759/rdgf.v24i3.2198 · Cited by 11
Ten Guidelines for Intelligent Systems Futures
Loi, D. (2018) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-02686-8_59 · Cited by 8
Optimization of Language Models by Word Computing
Chan, K., Im, S., & Zhang, Y. (2022) · International Conference on Graphics and Signal Processing (ICGSP) · doi:10.1145/3561518.3561525 · Cited by 7
“and King corpus [15], as well as the myPersonality corpus [36].”
BackgroundArtificial Intelligence: Positive or Negative Innovation
(2023) · International Journal of Emerging Trends in Engineering Research · doi:10.30534/ijeter/2023/031172023 · Cited by 7
Åsberg, C. (2024) · NORA - Nordic Journal of Feminist and Gender Research · doi:10.1080/08038740.2023.2294194 · Cited by 9
Facial Expression Recognition Using a Large Out-of-Context Dataset
Tran, E., Mayhew, M. B., Kim, H., Karande, P., & Kaplan, A. D. (2018) · IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) · doi:10.1109/wacvw.2018.00012 · Cited by 6
Inclusive AI in Recruiting. Multi-agent Systems Architecture for Ethical and Legal Auditing
Fernández, C., & Fernández, A. (2019) · Communications in computer and information science · doi:10.1007/978-3-030-24299-2_30 · Cited by 7
Innovation Management for Artificial Intelligence
Glauner, P. (2020) · Management for professionals · doi:10.1007/978-3-030-57642-4_1 · Cited by 7
Использование искусственного интеллекта в преступных целях: уголовно-правовая характеристика
Дремлюга, Р. (2021) · Азиатско-Тихоокеанский регион Экономика политика право · doi:10.24866/1813-3274/2021-3/153-165 · Cited by 7
Worrying Impact of Artificial Intelligence and Big Data Through the Prism of Recommender Systems
Bojić, L., Zarić, M., & Žikić, S. (2021) · Etnoantropološki problemi / Issues in Ethnology and Anthropology · doi:10.21301/eap.v16i3.13 · Cited by 7
Creepy Technologies and the Privacy Issues of Invasive Technologies
McWhorter, R. R., & Bennett, E. E. (2021) · IGI Global eBooks · doi:10.4018/978-1-7998-8954-0.ch083 · Cited by 8
ARTIFICIAL INTELLIGENCE AND INTERNATIONAL ARBITRATION
Malhoutra, A., & Ahmad, F. (2022) · Novos Estudos Jurídicos · doi:10.14210/nej.v27n2.p258-281 · Cited by 7
The Sins of the Parents Are to Be Laid Upon the Children: Biased Humans, Biased Data, Biased Models
Osborne, M. R., Omrani, A., & Dehghani, M. (2023) · Perspectives on Psychological Science · doi:10.1177/17456916231180099 · Cited by 9
“These models have appeared in research examining how facial-recognition software predicts users’ sexuality (Wang & Kosinski, 2018), what jobs people are likely to opt into (Song et al., 2022), how moral rhetoric online predicts future arrests during protests (Mooijman et al., 2018) and even in…”
Visual cues that predict intuitive risk perception in the case of HIV
Schmälzle, R., Hartung, F., Barth, A., Imhof, M. A., Kenter, A., Renner, B., & Schupp, H. T. (2019) · PLoS ONE · doi:10.1371/journal.pone.0211770 · Cited by 10
Adversariality in Machine Learning Systems: On Neural Networks and the Limits of Knowledge
Lepage-Richer, T. (2020) · Springer eBooks · doi:10.1007/978-3-030-56286-1_7 · Cited by 8
Social media and students’ privacy: What schools and districts should know
Rosenberg, J. M., Burchfield, M. A., Borchers, C., et al. (2021) · Phi Delta Kappan · doi:10.1177/00317217211051145 · Cited by 7
“We already know that facial recognition algorithms exist that use images to predict such characteristics as political identity (Kosinski, 2021) or sexual orientation of individuals 50 Kappan October 2021 G E TT Y I M A G E S Kappan V103 N2 51 from facial images (Wang & Kosinski, 2018).”
BackgroundANALYSIS OF PUBLIC DATA ON SOCIAL NETWORKS WITH IBM WATSON
Balogh, Z. (2018) · Acta Informatica Malaysia · doi:10.26480/aim.01.2018.10.11 · Cited by 7
Computer Vision Applications and their Ethical Risks in the Global South
Dufresne‐Camaro, C., Chevalier, F., & Ahmed, S. I. (2020) · Proceedings · doi:10.20380/gi2020.17 · Cited by 6
“Computer vision has also been leveraged for more controversial applications, such as sensitive attributes predictors for sexual orientation [80] and body weight [42], DeepFakes [78,81] and automated surveillance systems [53].”
BackgroundExcavating "Excavating AI": The Elephant in the GalleryPreprint
Lyons, M. J. (2020) · arXiv (Cornell University) · doi:10.17613/6f157-12p43 · Cited by 13
Intergroup Psychological Interventions: The Motivational ChallengePreprint
Landry, A. (2023) · doi:10.31234/osf.io/e5bcu · Cited by 7
Bendel, O. (2018) · National Conference on Artificial Intelligence · Cited by 6
Haut, K. G., Sen, T., Lomakin, D., & Hoque, M. E. (2022) · IEEE Transactions on Technology and Society · doi:10.1109/tts.2022.3172556 · Cited by 6
“For example, a recent AI system purports to be able to predict one’s sexual orientation from their facial features [53], [82].”
BackgroundFreedom of Thought: Who, What, and Why?
McCarthy‐Jones, S. (2021) · Springer eBooks · doi:10.1007/978-3-030-84494-3_2 · Cited by 6
Beyond the Binary --- Queering AI for an Inclusive Future
Liliequist, E., Tubella, A. A., Danielsson, K., et al. (2023) · interactions · doi:10.1145/3590141 · Cited by 6
“A highly debated study by Michal Kosinski and Yiland Wang [10] purports that classifiers can achieve a higher accuracy than humans in inferring sexual orientation from face images.”
BackgroundArtificial Intelligence and Fair Trial Rights
Molbæk-Steensig, H., & Quemy, A. (2023) · Oxford University Press eBooks · doi:10.1093/law/9780192882486.003.0018 · Cited by 7
Good AI for the Present of Humanity Democratizing AI Governance
Corrêa, N., & Fernandes de Oliveira, N. (2021) · AI Ethics Journal · doi:10.47289/aiej20210716-2 · Cited by 6
“A controversial study by Wang and Kosinski (2017), where the authors stated that ‘classification algorithms can infer sexual orientation from facial images’, caused a series of criticism from the LGBTQ+ community (Agüera y Arcas et al., 2018).”
BackgroundAlgorithmische Selektion und Privatheit
Dreyer, S., & Heldt, A. (2021) · Academia – ein Verlag in der Nomos Verlagsgesellschaft eBooks · doi:10.5771/9783896659378-117 · Cited by 6
Person, Human, Neither: The Dehumanization Potential of Automated Image Tagging
Barlas, P., Kyriakou, K., Kleanthous, S., & Otterbacher, J. (2021) · Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3461702.3462567 · Cited by 6
“Other types of computer vision applications have emerged that analyze people’s images (often their face) to infer an abstract quality about them – like personality [53], gender [23], sexual orientation [50], or emotional state [5] – even if it has not been tested whether a connection between the abstract quality and t…”
BackgroundNon-verbal signs of personality: Communicative meanings of facial expressions
Sharkov, F. I., Владимирович, С. В., & Феликсовна, К. О. (2022) · Вестник Российского университета дружбы народов. Серия: Социология · doi:10.22363/2313-2272-2022-22-2-387-403 · Cited by 6
Coding infant engagement in the Face-to-Face Still-Face paradigm using deep neural networks
Faltyn, M., Krzeczkowski, J. E., Cummings, M., Anwar, S., Zeng, T., Zahid, I., Ntow, K. O.-B., & Lieshout, R. J. V. (2023) · Infant Behavior and Development · doi:10.1016/j.infbeh.2023.101827 · Cited by 6
Artificial Intelligence and Freedom of Expression
Gregorio, G. D., & Dunn, P. (2023) · Oxford University Press eBooks · doi:10.1093/law/9780192882486.003.0005 · Cited by 6
Slavkovik, M., Stachl, C., Pitman, C., & Askonas, J. (2021) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3461702.3462626 · Cited by 9
“While the automated recognition of gender might seem banal, it can have a significant real-world impact once sexual orientation is predicted from the data people leave behind (Wang and Kosinski, 2018)1.”
BackgroundPolicy Gradient-Based Core Placement Optimization for Multichip Many-Core Systems
Myung, W., Lee, D., Song, C., Wang, G., & Ma, C. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3117878 · Cited by 7
“I N RECENT years, many algorithms based on deep neural networks have made remarkable progress in many application fields, such as computer vision [1], [2], speech recognition [3], playing games [4]–[6], robotics [7], natural language processing (NLP) [8], automation [9], and others [10].”
BackgroundLarge-scale GWAS reveals insights into the genetic architecture of same-sex sexual behavior
Ganna, A., Verweij, K. J. H., Nivard, M. G., Maier, R., Wedow, R., Busch, A. S., Abdellaoui, A., Guo, S., Sathirapongsasuti, J. F., Lichtenstein, P., et al. (2020) · Yearbook of pediatric endocrinology · doi:10.1530/ey.17.15.15 · Cited by 6
Good AI for the Present of Humanity Democratizing AI Governance
Corrêa, N. K., & Oliveira, N. F. D. (2021) · PhilPapers (PhilPapers Foundation) · Cited by 5
A cautionary note on predicting social judgments from faces with deep neural networksPreprint
Keleş, Ü., Lin, C., & Adolphs, R. (2021) · doi:10.31234/osf.io/t7hw4 · Cited by 6
Weidman, A. C., Sun, J., Vazire, S., et al. (2018) · doi:10.31234/osf.io/5w7c8 · Cited by 6
Dirty Data Labeled Dirt Cheap: Epistemic Injustice in Machine Learning SystemsPreprint
Hull, G. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4137697 · Cited by 6
AI and Its Importance in the Development of Luxury Consumption
Pawar, P., Anute, N., Tripathi, S., Ahluwalia, G. K., Kalshetti, P., & Selvakumar, P. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3693-8422-0.ch001 · Cited by 5
Understanding from Deep Learning Models in Context
Tamir, M., & Shech, E. (2022) · Routledge eBooks · doi:10.4324/9781003202905-28 · Cited by 6
Metadata
- Authors
- Wang, Y., & Kosinski, M.
- Affiliations
- Stanford University (US)
- ORCID
- Kosinski, Michal
- Venue
- Journal of Personality and Social Psychology
- Published
- 2018-02
- Volume
- 114
- Issue
- 2
- Pages
- 246-257
- ISSN
- 0022-3514 (ISSN-L)
- 1939-1315
- OpenAlex
- W2781265533
- Cited by
- 766
- References
- 66
- Citations by year
- 2026: 29 · 2025: 57 · 2024: 57 · 2023: 146 · 2022: 98 · 2021: 92 · 2020: 155 · 2019: 75 · 2018: 51
- FWCI
- 55.54
- Type
- journal article
- Language
- en
- Open access
- green
- doi.org
- Licence
- cc-by
- Locations
- Journal of Personality and Social Psychology (publishedVersion)
- PubMed (publishedVersion)
- PsyArXiv (OSF Preprints) (submittedVersion)
- OSF Preprints (OSF Preprints) (submittedVersion)
- PsyArXiv (OSF Preprints) (submittedVersion, cc-by)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Evolutionary Psychology and Human Behavior
- Keywords
- face images, deep neural network, gay men, gay women, facial morphology, privacy, computer vision, logistic regression
- MeSH terms
- Facial Recognition, Deep Learning, Adult, Face, Female, Humans, Male, Social Perception, Sexuality
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
