Cited by 778Face Recognition and PerceptionFace recognition and analysisEvolutionary Psychology and Human Behavior

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 access

Wang, Yilun, Kosinski, Michal·2018·Journal of Personality and Social Psychology

Published version of a preprint

Open this paper in CoChat — free account

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

Integrity signals

Registry

Crossref · 10.1037/pspa0000098

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

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?

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
Deep learning
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
Deep Face Recognition
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-citation
The 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
Facebook as a research tool for the social sciences: Opportunities, challenges, ethical considerations, and practical guidelines.
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

Flagged first, then by citations

Show all 66 references

13 of the 66 references couldn’t be matched to a known paper.

Sorted by citations

Items 101–200 of 778

Digital Threats to Democratic Elections: How Foreign Actors Use Digital Techniques to Undermine DemocracyPreprint
Tenove, C., Buffie, J., McKay, S., & Moscrop, D. (2018) · SSRN Electronic Journal · doi:10.2139/ssrn.3235819 · Cited by 50
AI as super-controversy: Eliciting AI and society controversies with an extended expert community in the UK
Marres, N., Castelle, M., Gobbo, B., Poletti, C., & Tripp, J. (2024) · Big Data & Society · doi:10.1177/20539517241255103 · Cited by 56
Facial Recognition Technology: A Primer for Plastic Surgeons
Zuo, K. J., Saun, T. J., & Forrest, C. R. (2019) · Plastic & Reconstructive Surgery · doi:10.1097/prs.0000000000005673 · Cited by 36
PE-MIU: A Training-Free Privacy-Enhancing Face Recognition Approach Based on Minimum Information Units
Terhörst, P., Riehl, K., Damer, N., et al. (2020) · IEEE Access · doi:10.1109/access.2020.2994960 · Cited by 36
Gender Classification and Bias Mitigation in Facial Images
Wu, W., Protopapas, P., Yang, Z., & Michalatos, P. (2020) · ACM Web Science Conference (WebSci) · doi:10.1145/3394231.3397900 · Cited by 36
Ethical Challenges Posed by Big Data.
Howe, E. G., & Elenberg, F. (2020) · PubMed Central · Cited by 37
Computer Vision and Conflicting Values: Describing People with Automated Alt Text
Hanley, M., Barocas, S., Levy, K., Azenkot, S., & Nissenbaum, H. (2021) · Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3461702.3462620 · Cited by 38
“Other scholars have remarked on the dangers of using computer vision to draw inferences about personal characteristics that are not visually evident [70, 62], including sexual orientation and criminality [67, 29].”
Background
Algorithms, ontology, and social progress
Iliadis, A. (2018) · Global Media and Communication · doi:10.1177/1742766518776688 · Cited by 33
“Recent studies have even reported that algorithms are more accurate than humans at detecting sexual orientation (Wang and Kosinski, 2017), disguised faces (Singh et al., 2017) and criminality using face images (Wu and Zhang, 2016).”
Background
Gay Voice: Stable Marker of Sexual Orientation or Flexible Communication Device?
Daniele, M., Fasoli, F., António, R., Sulpizio, S., & Maass, A. (2020) · Archives of Sexual Behavior · doi:10.1007/s10508-020-01771-2 · Cited by 34
Inteligência Artificial em Cardiologia: Conceitos, Ferramentas e Desafios – “Quem Corre é o Cavalo, Você Precisa ser o Jóquei”
Filho, E. M. D. S., Fernandes, F. D. A., Soares, C. L. D. A., Seixas, F. L., dos Santos, A. A. S. M. D., Gismondi, R. A., Mesquita, E. T., & Mesquita, C. T. (2019) · Arquivos Brasileiros de Cardiologia · doi:10.36660/abc.20180431 · Cited by 47
Speech Acoustic Features: A Comparison of Gay Men, Heterosexual Men, and Heterosexual Women
Suire, A., Tognetti, A., Durand, V., et al. (2020) · Archives of Sexual Behavior · doi:10.1007/s10508-020-01665-3 · Cited by 33
Honest-but-Curious Nets: Sensitive Attributes of Private Inputs Can Be Secretly Coded into the Classifiers' Outputs
Malekzadeh, M., Borovykh, A., & Gündüz, D. (2021) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3460120.3484533 · Cited by 38
Transferable face image privacy protection based on federated learning and ensemble models
Yang, J., Liu, J., Han, R., & Wu, J. (2021) · Complex & Intelligent Systems · doi:10.1007/s40747-021-00399-6 · Cited by 32
Deep learning analysis and age prediction from shoeprints
Hassan, M. A., Wang, Y., Wang, D., Li, D., Liang, Y., Zhou, Y., & Xu, D. H. (2021) · Forensic Science International · doi:10.1016/j.forsciint.2021.110987 · Cited by 32
“In contrast to age predictions based on other human traits including facial images [28], [29] brain MRI and EEG [30], [31], [32], DNA [33] and gait patterns [34], [35], age prediction from shoeprints is more challenging due to image noises, varying patterns, and shoe designs.”
ResultSelf-citation
Incremental and Approximate Inference for Faster Occlusion-based Deep CNN Explanations
Nakandala, S., Kumar, A., & Papakonstantinou, Y. (2019) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3299869.3319874 · Cited by 30
“Thus, there is growing interest in adopting deep CNNs in various application domains, including healthcare [2, 3], agri-culture [4], security [5], and sociology [6].”
Background
SoK: Differential Privacy as a Causal Property
Tschantz, M. C., Sen, S., & Datta, A. (2020) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp40000.2020.00012 · Cited by 36
“…probability where person i does not provide data or provides fake data is where To gain intuition about these results, let us consider the findings of Wang and Kosinski [47], which show the possibility of training a neural network to predict a person’s sexual orientation from a photo of their face.”
Background
Digital Facial Anthropometry: Application and Implementation
Kukharev, G., & Kaziyeva, N. (2020) · Pattern Recognition and Image Analysis · doi:10.1134/s1054661820030141 · Cited by 31
“For example, Deep NN determined two psychological types according to eye-brows, eyes, and mouth micro-motions and according to facial expression [79, 80].”
Background
Artificial Intelligence and the Practice of Radiology: An Alternative View
Schier, R. (2018) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2018.03.046 · Cited by 31
We, the Data: Human Rights in the Digital Age
Wong, W. H. (2023) · doi:10.7551/mitpress/14703.001.0001 · Cited by 31
Inductive Risk, Understanding, and Opaque Machine Learning Models
Sullivan, E. (2022) · Philosophy of Science · doi:10.1017/psa.2022.62 · Cited by 29
“For example, Wachter et al. (2018) provides an interpretability method that involves counterfactual extraction specifically, providing a counterfactual explanation along the lines of a “ what if things had been different?”
Method
Large language models know how the personality of public figures is perceived by the general public
Cao, X., & Kosiński, M. (2024) · Scientific Reports · doi:10.1038/s41598-024-57271-z · Cited by 35
Self-citation
Is Generative AI Increasing the Risk for Technology‐Mediated Trauma Among Vulnerable Populations?
Abdulai, A. (2024) · Nursing Inquiry · doi:10.1111/nin.12686 · Cited by 36
The nexus of artificial intelligence and entrepreneurship research: Bibliometric analysis
Mumi, A., Ngammoh, N., & Suwanpakdee, A. (2025) · Sustainable Futures · doi:10.1016/j.sftr.2025.100688 · Cited by 35
The locus of legitimate interpretation in Big Data sciences: Lessons for computational social science from -omic biology and high-energy physics
Bartlett, A., Lewis, J., Reyes-Galindo, L. I., & Stephens, N. (2018) · Big Data & Society · doi:10.1177/2053951718768831 · Cited by 32
Biometric Mirror
Wouters, N., Kelly, R., Velloso, E., Wolf, K., Ferdous, H. S., Newn, J., Joukhadar, Z., & Vetere, F. (2019) · Designing Interactive Systems Conference · doi:10.1145/3322276.3322304 · Cited by 32
“Embody Concerns: Recent reports and academic work highlight how facial analysis models are used to infer sensitive data from facial expressions, including perceived traits and human behaviors [32, 36, 49, 86].”
Background
Usable and secure? User perception of four authentication methods for mobile banking
Kružíková, A., Knapová, L., Šmahel, D., Dědková, L., & Matyáš, V. (2022) · Computers & Security · doi:10.1016/j.cose.2022.102603 · Cited by 28
Artificial Intelligence Crime: An Interdisciplinary Analysis of Foreseeable Threats and Solutions
King, T. C., Aggarwal, N., Taddeo, M., & Floridi, L. (2021) · Philosophical studies series · doi:10.1007/978-3-030-81907-1_13 · Cited by 35
Movie genre preference prediction using machine learning for customer-based information
Wang, H., & Zhang, H. (2018) · IEEE Annual Computing and Communication Workshop and Conference (CCWC) · doi:10.1109/ccwc.2018.8301647 · Cited by 28
“The interested reader is referred to [31], [32] for a more complete description.”
Background
Big Data, Big Concerns: Ethics in the Digital Age
Jurkiewicz, C. L. (2018) · Public Integrity · doi:10.1080/10999922.2018.1448218 · Cited by 30
From privacy to anti-discrimination in times of machine learning
Hagendorff, T. (2019) · Ethics and Information Technology · doi:10.1007/s10676-019-09510-5 · Cited by 30
“For example, a study by Kosinski and Wang (2017) showed that computers can be taught to recognize the sexual orientation of people—the study was later successfully replicated (Leuner 2019)—or to classify people into “criminals” and “non-criminals” (Wu and Zhang 2016) on the basis of profile…”
Background
Ethical Considerations of Digital Phenotyping from the Perspective of a Healthcare Practitioner
Dagum, P., & Montag, C. (2019) · Studies in neuroscience, psychology and behavioral economics · doi:10.1007/978-3-030-31620-4_2 · Cited by 30
RETRACTED ARTICLE: Criminal tendency detection from facial images and the gender bias effect
Hashemi, M., & Hall, M. (2020) · Journal Of Big Data · doi:10.1186/s40537-019-0282-4 · Cited by 28
Retracted
Deep learning-based facial image analysis in medical research: a systematic review protocol
Su, Z., Liang, B., Shi, F., Gelfond, J., Šegalo, S., Wang, J., Jia, P., & Hao, X. (2021) · BMJ Open · doi:10.1136/bmjopen-2020-047549 · Cited by 34
“5 7 10 34–36 Comparing clinical and deep learning evaluations of microdeletion syndrome facial phenotypes, researchers found that deep learning outperformed clinical evaluations in terms of sensitivity and specificity by 96%.”
Background
A social network of crime: A review of the use of social networks for crime and the detection of crime
Drury, B., Drury, S. M., Rahman, M. A., & Ullah, I. (2022) · Online Social Networks and Media · doi:10.1016/j.osnem.2022.100211 · Cited by 29
The rise of artificial intelligence, the fall of human wellbeing?
Zhao, Y., Yin, D., Wang, L., & Yu, Y. (2023) · International Journal of Social Welfare · doi:10.1111/ijsw.12586 · Cited by 29
The Attention Economy and the Impact of Artificial Intelligence
Baeza-Yates, R., & Fayyad, U. (2021) · Perspectives on Digital Humanism · doi:10.1007/978-3-030-86144-5_18 · Cited by 8
“That includes discrimination (e.g., persons, products, businesses), phrenology (e.g., prediction of personality traits based in facial biometrics (Wang and Kosinski 2018)), data used without consent (e.g., faces scrapped from the Web used to train facial recognition systems (Raji and Fried 2020)),…”
Method
Simulation vs. Understanding: A Tension, in Quantum Chemistry and Beyond. Part B. The March of Simulation, for Better or Worse
Hoffmann, R., & Malrieu, J. (2019) · Angewandte Chemie International Edition · doi:10.1002/anie.201910283 · Cited by 26
Deep neural networks and image classification in biological vision
Leek, E. C., Leonardis, A., & Heinke, D. G. (2022) · Vision Research · doi:10.1016/j.visres.2022.108058 · Cited by 29
What makes a market transaction morally repugnant?
Leuker, C., Samartzidis, L., & Hertwig, R. (2021) · Cognition · doi:10.1016/j.cognition.2021.104644 · Cited by 23
The Three Harms of Gendered Technology
Wang, L. (2020) · AJIS. Australasian journal of information systems/AJIS. Australian journal of information systems/Australian journal of information systems · doi:10.3127/ajis.v24i0.2799 · Cited by 29
“For example, an AI algorithm has been developed to detect a person’s sexual orientation (Wang and Kosinski, 2017) which has privacy implications.”
Method
kgbench: A Collection of Knowledge Graph Datasets for Evaluating Relational and Multimodal Machine Learning
Bloem, P., Wilcke, X., Berkel, L. V., & de Boer, V. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-77385-4_37 · Cited by 31
“Here, a comparable case is [33,14], where a classifier was built to predict sexual orientation.”
Background
Facial first impressions form two clusters representing approach-avoidance
Jones, A. L., & Kramer, R. S. S. (2021) · Cognitive Psychology · doi:10.1016/j.cogpsych.2021.101387 · Cited by 26
Generative AI Ethical Considerations and Discriminatory Biases on Diverse Students Within the Classroom
Salazar, L. R., Peeples, S., & Brooks, M. E. (2024) · Advances in educational technologies and instructional design book series · doi:10.4018/979-8-3693-0831-8.ch010 · Cited by 26
Confident Privacy Decision-Making in IoT Environments
Lee, H., & Kobsa, A. (2019) · ACM Transactions on Computer-Human Interaction · doi:10.1145/3364223 · Cited by 23
“Informing users in detail about sensor data collection and analysis is important since due to the rapid advancements in data mining and machine learning (ML), all kinds of data can be used to infer potentially sensitive personal information (e.g., sexual orientation [73]).”
Background
Artificial Intelligence for marketing plan: the case for e-marketing companies
Fayed, A. E. (2021) · Marketing and Management of Innovations · doi:10.21272/mmi.2021.1-07 · Cited by 32
Internet Users’ Valuation of Enhanced Data Protection on Social Media: Which Aspects of Privacy Are Worth the Most?
Mahmoodi, J., Čurdová, J., Henking, C., Kunz, M. C., Matić, K., Mohr, P., & Vovko, M. (2018) · Frontiers in Psychology · doi:10.3389/fpsyg.2018.01516 · Cited by 24
Platform Biometrics
Crampton, J. W. (2019) · Surveillance & Society · doi:10.24908/ss.v17i1/2.13111 · Cited by 26
Incremental and Approximate Computations for Accelerating Deep CNN Inference
Nakandala, S., Nagrecha, K., Kumar, A., & Papakonstantinou, Y. (2020) · ACM Transactions on Database Systems · doi:10.1145/3397461 · Cited by 22
“Thus, there is growing adoption of deep CNNs in numerous application domains such as healthcare [2, 3], agriculture [4], image search and recommendation systems [5], species monitoring [6], security [7], and sociology [8].”
Background
What People Think AI Should Infer From Faces
Engelmann, S., Ullstein, C., Papakyriakopoulos, O., & Großklags, J. (2022) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3531146.3533080 · Cited by 28
Dirty Data Labeled Dirt Cheap: Epistemic Injustice in Machine Learning Systems
Hull, G. (2022) · Ethics and Information Technology · doi:10.1007/s10676-023-09712-y · Cited by 32
“In 2018, Kosinski and Yulin Wang published a paper claiming that an AI could be trained to do better than a human in predicting sexual orientation from a facial image (Wang and Kosinski, 2018).”
Background
Frontiers in Service Science: The Management of Data Analytics Services: New Challenges and Future Directions
de Véricourt, F., & Perakis, G. (2020) · Service Science · doi:10.1287/serv.2020.0262 · Cited by 20
“For instance, ML algorithm are better than humans at predicting the sexual orientation of individuals through facial recognition (Wang and Kosinski 2018).”
Background
Utilizing large language models for gastroenterology research: a conceptual framework
Berry, P., Dhanakshirur, R. R., & Khanna, S. (2025) · Therapeutic Advances in Gastroenterology · doi:10.1177/17562848251328577 · Cited by 26
Künstliche Intelligenz in der Psychiatrie – ein Überblick
Meyer‐Lindenberg, A. (2018) · Der Nervenarzt · doi:10.1007/s00115-018-0557-6 · Cited by 22
“Ein aktuelles Beispiel sind Daten, die zeigen, dass sexuelle Orientierung aus Photos, wie sie im Internet anzutreffen sind, von NN mit deutlich größerer Genauigkeit geschätzt werden kann als von menschlichen Experten [38].”
Influential
Generating and Restoring Private Face Images for Internet of Vehicles Based on Semantic Features and Adversarial Examples
Yang, J., Liu, J., Han, R., & Wu, J. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3102266 · Cited by 21
“…face images are not processed in any way when face images are published or obtained by criminals during transmission, criminals may use these features to obtain users’ personal, private data via data-mining and other techniques, misappro-priating users’ identities for criminal purposes [4], [5].”
Background
Body stakes: an existential ethics of care in living with biometrics and AI
Lagerkvist, A., Tudor, M., Smolicki, J., Ess, C., Lundström, J. E., & Rogg, M. (2022) · AI & Society · doi:10.1007/s00146-022-01550-8 · Cited by 27
Good Intentions, Bad Inventions: How Employees Judge Pervasive Technologies in the Workplace
Constantinides, M., & Quercia, D. (2022) · IEEE Pervasive Computing · doi:10.1109/mprv.2022.3217408 · Cited by 21
“Tracking facial expressions (even beyond the nefarious uses—of dubious effectiveness—of inferring political orientation or sexual preferences from faces [19]) is possible and could be done in seamless ways (e.”
Background
(Not) hearing happiness: Predicting fluctuations in happy mood from acoustic cues using machine learning.
Weidman, A. C., Sun, J., Vazire, S., Quoidbach, J., Ungar, L., & Dunn, E. W. (2019) · Emotion · doi:10.1037/emo0000571 · Cited by 24
“Machine learning algorithms have recently been used to gain insight into difficult prediction-related psychological research questions (e.g., Joel, Eastwick, & Finkel, 2017; Laukka et al., 2014; Wang & Kosinski, 2018).”
Method
Research on manufacturing service combination optimization based on neural network and multi-attribute decision making
Yang, M. N., Zhu, H. T., & Guo, K. (2019) · Neural Computing and Applications · doi:10.1007/s00521-019-04241-6 · Cited by 21
Assessing Trust Versus Reliance for Technology Platforms by Systematic Literature Review
Deley, T., & Dubois, E. (2020) · Social Media + Society · doi:10.1177/2056305120913883 · Cited by 24
Current cancer driver variant predictors learn to recognize driver genes instead of functional variants
Raimondi, D., Passemiers, A., Fariselli, P., & Moreau, Y. (2021) · BMC Biology · doi:10.1186/s12915-020-00930-0 · Cited by 22
Context-Contingent Privacy Concerns and Exploration of the Privacy Paradox in the Age of AI, Augmented Reality, Big Data, and the Internet of Things: Systematic Review
Herriger, C., Merlo, O., Eisingerich, A. B., & Arigayota, A. (2025) · Journal of Medical Internet Research · doi:10.2196/71951 · Cited by 34
“An example of this is how AI can predict sexual orientation with up to 91% accuracy from just 5 images of a person’s face [64], which Nissenbaum [40] refers to as the data food chain (ie, data of higher order are inferred from lower-order data).”
Consent in the Age of AR: Investigating The Comfort With Displaying Personal Information in Augmented Reality
Rixen, J. O., Colley, M., Askari, A., Gugenheimer, J., & Rukzio, E. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3502140 · Cited by 21
“AI also has started to outperform humans in a multitude of other areas like estimating a person’s age [19], sexual orientation, or race [20] through visual clues [63].”
Background
The value of protecting privacy
Santanen, E. L. (2018) · Business Horizons · doi:10.1016/j.bushor.2018.04.004 · Cited by 21
Polar Similars: Using Massive Mobile Dating Data to Predict Synchronization and Similarity in Dating Preferences
Levy, J., Markell, D., & Cerf, M. (2019) · Frontiers in Psychology · doi:10.3389/fpsyg.2019.02010 · Cited by 22
“Recent works in computer vision and psychology suggest that some personality attributes can in fact be gathered merely from the visual imagery (Cerf et al., 2008; Wang and Kosinski, 2018).”
Background
Forbidden knowledge in machine learning reflections on the limits of research and publication
Hagendorff, T. (2020) · AI & Society · doi:10.1007/s00146-020-01045-4 · Cited by 25
“” (Kosinski and Wang 2018) To prevent this threat to a certain”
BackgroundInfluential
Tracking Fluctuations in Psychological States using Social Media Language: A Case Study of Weekly Emotion
Eichstaedt, J. C., & Weidman, A. C. (2020) · European Journal of Personality · doi:10.1002/per.2261 · Cited by 22
“Given the sensitive nature of big data analyses and the consent process, researchers still at times face understandable backlash over the potential of these methods to derive sensitive personal information (e.g. Kramer, Guillory, & Hancock, 2014; Wang & Kosinski, 2018).”
Background
Putative ratios of facial attractiveness in a deep neural network
Tong, S., Liang, X., Kumada, T., & Iwaki, S. (2020) · Vision Research · doi:10.1016/j.visres.2020.10.001 · Cited by 20
CEO Stress, Aging, and Death
Borgschulte, M., Guenzel, M., Liu, C., & Malmendier, U. M. (2021) · National Bureau of Economic Research · doi:10.3386/w28550 · Cited by 18
Accuracy-Efficiency Trade-Offs and Accountability in Distributed ML Systems
Cooper, A. F., Levy, K. E. C., & De, C. (2021) · ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO) · doi:10.1145/3465416.3483289 · Cited by 21
The Future of Hybrid Meetings
Constantinides, M., & Quercia, D. (2022) · Annual Symposium on Human-Computer Interaction for Work (CHIWORK) · doi:10.1145/3533406.3533415 · Cited by 25
“Tracking facial expressions (even beyond the nefarious uses - of dubious effectiveness - of inferring political orientation or sexual preferences [43]) is possible and could be done in seamless ways (e.”
Background
Using deep learning to predict ideology from facial photographs: expressions, beauty, and extra-facial information
Rasmussen, S. H. R., Ludeke, S. G., & Klemmensen, R. (2023) · Scientific Reports · doi:10.1038/s41598-023-31796-1 · Cited by 19
Bias in Facial Recognition Technologies Used by Law Enforcement: Understanding the Causes and Searching for a Way Out
Limantė, A. (2023) · Nordic Journal of Human Rights · doi:10.1080/18918131.2023.2277581 · Cited by 30
Visual Data and the Law
Brayne, S., Levy, K. E. C., & Newell, B. C. (2018) · Law & Social Inquiry · doi:10.1111/lsi.12373 · Cited by 19
A Novel Statistical Feature Analysis-Based Global and Local Method for Face Recognition
Talab, M. A., Awang, S., & Ansari, M. D. (2020) · International Journal of Optics · doi:10.1155/2020/4967034 · Cited by 19
“*is method dynamically groups neurons to a higher order; hence, it is robust against environment variations [70, 71].”
Method
Decoding depressive disorder using computer vision
Singh, J., & Goyal, G. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-10128-9 · Cited by 21
Hidden depths: The effects of extrinsic data collection on consumer insurance contracts
Bednarz, Z., & Manwaring, K. (2022) · Computer law & security review · doi:10.1016/j.clsr.2022.105667 · Cited by 19
Identifying psychological indexes based on social media data: A machine learning method
Su, Y., Liu, M., Nan, Z. H. A. O., Liu, X., & Zhu, T. (2021) · Advances in Psychological Science · doi:10.3724/sp.j.1042.2021.00571 · Cited by 19
Representativeness and face-ism: Gender bias in image search
Ulloa, R., Richter, A. C., Makhortykh, M., Urman, A., & Kacperski, C. (2022) · New Media & Society · doi:10.1177/14614448221100699 · Cited by 18
“Critical evaluation as to the consequences of such technologies has already been undertaken with regards to sexual orientation; they are often a threat to individuals’ privacy and safety (e.g. Raji et al., 2020; Wang and Kosinski, 2018).”
Background
Identification of the Facial Features of Patients With Cancer: A Deep Learning–Based Pilot Study
Liang, B., Yang, N., He, G., Huang, P., & Yang, Y. (2020) · Journal of Medical Internet Research · doi:10.2196/17234 · Cited by 19
“Wang et al [19] found that the DL technique outperformed humans at detecting sexual orientation based on facial images.”
Background
Revolution 4.0 and its implications for consumer behaviour
Giza, W., & Wilk, B. (2021) · Entrepreneurial Business and Economics Review · doi:10.15678/eber.2021.090412 · Cited by 18
Editorial overview: Big data in the behavioral sciences
Kosiński, M., & Behrend, T. S. (2017) · Current Opinion in Behavioral Sciences · doi:10.1016/j.cobeha.2017.11.007 · Cited by 18
Self-citation
Synthetic Realities in the Digital Age: Navigating the Opportunities and Challenges of AI-Generated ContentPreprint
Wang, Y. (2023) · doi:10.36227/techrxiv.23968311.v1 · Cited by 23
Self-citation
Is there a civic duty to support medical AI development by sharing electronic health records?
Müller, S. (2022) · BMC Medical Ethics · doi:10.1186/s12910-022-00871-z · Cited by 17
“For example, Wang and Kosinski built a deep neural network that used a data-base of 35,326 facial images of self-reported homosexual men and women living in the US to learn how to recognise facial expressions that are characteristic of gay men and women [74].”
Background
A call for an ethical framework when using social media data for artificial intelligence applications in public health research
Gilbert, J.-P., Ng, V., Niu, J., & Rees, E. E. (2020) · Canada Communicable Disease Report · doi:10.14745/ccdr.v46i06a03 · Cited by 18
“(22) reported being able to identify people’s sexual orientation by processing pictures of people from a dating website.”
Background
Unsupervised Enhancement of Soft-biometric Privacy with Negative Face RecognitionPreprint
Terhörst, P., Huber, M., Damer, N., Kirchbuchner, F., & Kuijper, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.09181 · Cited by 19
“With the use of soft-biometric estimators, information about gender, age, ethnicity, sexual orientation or the health status can be obtained [6] [7].”
Method
Counterfactuals and Causability in Explainable Artificial Intelligence: Theory, Algorithms, and ApplicationsPreprint
Chou, Y.-L., Moreira, C., Bruza, P., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.04244 · Cited by 17
“Other studies revealed gender and racial biases in automated facial analysis algorithms made available by commercial companies [11], gender biases in textual predictive models [14, 15, 16] or even more discriminatory topics such as facial features according to sexual orientation [17].”
Background
Understanding the Research Landscape of Deep Learning in Biomedical Science: Scientometric Analysis
Nam, S., Kim, D., Jung, W., & Zhu, Y. (2022) · Journal of Medical Internet Research · doi:10.2196/28114 · Cited by 21
“VGG-Face Facial images of hetero and homosexual Classification A39 [82]”
Method
The paradoxical transparency of opaque machine learning
Lo, F. T. H. (2022) · AI & Society · doi:10.1007/s00146-022-01616-7 · Cited by 20
Facial recognition technology and human raters can predict political orientation from images of expressionless faces even when controlling for demographics and self-presentation.
Kosiński, M., Khambatta, P., & Wang, Y. (2024) · American Psychologist · doi:10.1037/amp0001295 · Cited by 14
“…infer sensitive traits such as political orientation (Joo et al., 2015; Kosinski, 2021; Rasmussen et al., 2023), personality (Kachur et al., 2020; Moreno-Armendariz et al., 2020; Segalin et al., 2017; Xu et al., 2021), and sexual orientation (Leuner, 2019; D. Wang, 2022; Y. Wang & Kosinski, 2018).”
BackgroundSelf-citation
The reanimation of pseudoscience in machine learning and its ethical repercussions
Andrews, M., Smart, A., & Birhane, A. (2024) · Patterns · doi:10.1016/j.patter.2024.101027 · Cited by 18
“By relying on ID photographs, the only factors controlled for were those speciﬁc to camera, lighting conditions, and at least in principle, the positioning of the head, factors that, e.g., Kabir et al. 8 and Wang and Kosinski 20,21 do not control for.”
Influential
Using advanced machine learning algorithms to predict academic major completion: A cross-sectional study
Kordbagheri, A., Kordbagheri, M., Tayim, N., Fakhrou, A. A., & Davoudi, M. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109372 · Cited by 19
Facial Image Privacy Protection Based on Principal Components of Adversarial Segmented Image Blocks
Yang, J., Liu, J., & Wu, J. (2020) · IEEE Access · doi:10.1109/access.2020.2999449 · Cited by 16
Helping Users Tackle Algorithmic Threats on Social Media: A Multimedia Research Agenda
von der Weth, C., Abdul, A., Fan, S., & Kankanhalli, M. (2020) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3394171.3414692 · Cited by 17
“[94] [111] [215] [36] income/occupation [97] [130] [89] [160] personality traits [9] [66] [180] [124] [150] [68] [98] [188] [4] [78] emotion & mood [13] [147] [122] [24] [127] sexual orientation [204] [100] health (social) well-being [39 ] [174 134 187 31] [62] [73] mental health [20 ] [50 82 19 34 56] [43 ] [189…”
Ética, Inteligência Artificial e Cardiologia
Filho, E. M. D. S., Fernandes, F. D. A., Pereira, N. C. D. A., Mesquita, C. T., & Gismondi, R. A. (2020) · Arquivos Brasileiros de Cardiologia · doi:10.36660/abc.20200143 · Cited by 15
Item Response Theory -- A Statistical Framework for Educational and Psychological Measurement
Chen, Y., Li, X., Liu, J., et al. (2021) · London School of Economics and Political Science Research Online (London School of Economics and Political Science) · doi:10.48550/arxiv.2108.08604 · Cited by 15
“…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).”
Optimising Emotions, Incubating Falsehoods
Bakir, V., & McStay, A. (2022) · Springer eBooks · doi:10.1007/978-3-031-13551-4 · Cited by 20
Epistemic Therapy for Bias in Automated Decision-Making
Gilbert, T. K., & Mintz, Y. D. (2019) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3306618.3314294 · Cited by 16
“This moral disposition is conveyed by the recent work of Corbett-Davies and Goel [11] on fairness measures, and Wang and Kosinski’s [34] work on detecting sexual orientation with pictures .”
BackgroundMethodInfluential
Smartphones within Psychological Science
Ellis, D. A. (2020) · Cambridge University Press eBooks · doi:10.1017/9781108671408 · Cited by 14
Digitale Phänotypisierung in der Psychologie – ein Quantensprung in der psychologischen Forschung?
Baumeister, H., Garatva, P., Pryss, R. C., Ropinski, T., & Montag, C. (2022) · Psychologische Rundschau · doi:10.1026/0033-3042/a000609 · Cited by 15
A Cautionary Note on Predicting Social Judgments from Faces with Deep Neural Networks
Keleş, Ü., Lin, C., & Adolphs, R. (2021) · Affective Science · doi:10.1007/s42761-021-00075-5 · Cited by 18

Authors: Wang, Y., & Kosinski, M.
Affiliations: Stanford University (US)
ORCID: Kosinski, Michal
Venue: Journal of Personality and Social Psychology
Publisher: American Psychological Association (APA)
Published: 2018-02
Volume: 114
Issue: 2
Pages: 246-257
ISSN: 0022-3514 (ISSN-L)
1939-1315
DOI: 10.1037/pspa0000098
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

Deep neural networks are more accurate than humans at detecting sexual orientation from facial images.

Abstract

Overview

Integrity signals

Verification

Claims the paper makes

Claims other papers make about this one

Citations

Cites

Cited by

Related

Metadata

Record

About these checks