Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing
Open accessThe DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Although essential to revealing biased performance, well intentioned attempts at algorithmic auditing can have effects that may harm the very populations these measures are meant to protect. This concern is even more salient while auditing biometric systems such as facial recognition, where the data is sensitive and the technology is often used in ethically questionable manners. We demonstrate a set of fiveethical concerns in the particular case of auditing commercial facial processing technology, highlighting additional design considerations and ethical tensions the auditor needs to be aware of so as not exacerbate or complement the harms propagated by the audited system. We go further to provide tangible illustrations of these concerns, and conclude by reflecting on what these concerns mean for the role of the algorithmic audit and the fundamental product limitations they reveal.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“As we have shown in this paper, it may be difcult (if not impossible) for AIaaS providers to exhaustively test and validate their services as free from all kinds of bias in all kinds of contexts [127,136].”
From Out of Context: Investigating the Bias and Fairness Concerns of “Artificial Intelligence as a Service” (2023) · cites this paper as [127]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“While there is a growing interest in auditing algorithmic systems [125,126,128,138,154], this feld generally lacks defned standards and records could be incomplete or not relating to information needed for accountability purposes [15,127].”
From Out of Context: Investigating the Bias and Fairness Concerns of “Artificial Intelligence as a Service” (2023) · cites this paper as [127]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“It is well acknowledged that machine learning models, such as those ofered by the API providers, are prone to exhibiting various kinds of discriminatory behaviours, including diferent error rates across demographic groups [22,126,127], stereotyping [64,163] or minority groups under-representation [134,137].”
From Out of Context: Investigating the Bias and Fairness Concerns of “Artificial Intelligence as a Service” (2023) · cites this paper as [127]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“However, even in instances when the algorithmic systems are themselves disclosed, it remains difficult to anticipate which of these deployments are linked to known incidents of harms, and thus challenging to prioritize cases or pursue targets proactively [64].”
From Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance (2022) · cites this paper as [64]
Experimental — accuracy not yet measuredIn the citing paper
“We note that just as it is recognised that transparency won't necessary solve algorithmic accountability issues [9,85,127], transparency will not by itself address all AIaaS concerns.”
From Out of Context: Investigating the Bias and Fairness Concerns of “Artificial Intelligence as a Service” (2023) · cites this paper as [127]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Crenshaw, K. (2018) · Feminist Legal Theory · doi:10.4324/9780429500480-5 · Cited by 9,510
Crenshaw, K. (1998) · Feminism And Politics · doi:10.1093/oso/9780198782063.003.0016 · Cited by 5,672
(2018) · Cited by 3,510
Self-citationReview of “Race After Technology: Abolitionist Tools for the New Jim Code”
Jipguep-Akhtar, M. (2020) · Social Forces · doi:10.1093/sf/soz162 · Cited by 1,473 · cited 1×
“The reality of FPT deployments, however, reveals that many of these uses are quite vulnerable to abuse, especially when used for surveillance and coupled with predatory data collection practices that, intentionally or unintentionally, discriminate against marginalized groups [2].”
BackgroundDeep Expectation of Real and Apparent Age from a Single Image Without Facial Landmarks
Rothe, R., Timofte, R., & Van Gool, L. (2018) · International Journal of Computer Vision · doi:10.1007/s11263-016-0940-3 · Cited by 896 · cited 4×
“oday, and none that enables an intersectional black box audit evaluation for such a wide range of tasks. To address this gap, we develop a benchmark called CelebSET, a subset of the IMDB-WIKI dataset [40] that includes 80 celebrity identities–20 of the most photographed celebrities from each of the subgroups‘darkermal…”
BackgroundMethodInfluentialRaji, I. D., & Buolamwini, J. (2019) · Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3306618.3314244 · Cited by 651
Self-citationWhere fairness fails: data, algorithms, and the limits of antidiscrimination discourse
Hoffmann, A. L. (2019) · Information, Communication & Society · doi:10.1080/1369118x.2019.1573912 · Cited by 535
The Misgendering Machines: Trans/HCI Implications of Automatic Gender Recognition
Keyes, O. (2018) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3274357 · Cited by 505
Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness
Kearns, M., Neel, S., Roth, A., & Wu, Z. S. (2017) · arXiv · doi:10.48550/arxiv.1711.05144 · Cited by 469 · cited 1×
“This is an observation of the fairness gerrymandering effect [28] - where optimizing for fairness on one axis can compromise fairness in another.”
BackgroundFitzpatrick skin typing: applications in dermatology
Sachdeva, S. (2009) · Indian Journal of Dermatology, Venereology and Leprology · doi:10.4103/0378-6323.45238 · Cited by 311 · cited 1×
“…is approximated by taking a subset of the images of celebrities tagged with the crawled ‘Black’ label from our crawled dataset, and selecting subjects within the 3 to 6 range of the Fitzpatrick skin type scale [41] using the reported Fitzpatrick labels of celebrity references for guidance.”
Method
Research works printed but not linked: 14
- Amazon. 2019. Amazon Rekognition FAQs. Retrieved Oct 31, 2019 from https://aws.amazon.com/rekognition/faqs/
- Sebastian Benthall and Bruce D. Haynes. 2019. Racial Categories in Machine Learning. In Proc. of the Conference on Fairness, Accountability, and Transparency (FAT). 10.
- Clarifai. 2019. Custom Face Recognition. Retrieved Oct 31, 2019 from https: //www.clarifai.com/custom-face-recognition
- Lord Clement-Jones. 2019. Automated Facial Recognition Technology (Moratorium and Review) Bill [HL] 2019--20. https://services.parliament.uk/bills/2019- 20/automatedfacialrecognitiontechnologymoratoriumandreview.html
- Craig R Ehlen and Robert B Welker. 1996. Procedural fairness in the peer and quality review programs. Auditing 15, 1 (1996), 38.
- Face++. 2019. Face Attributes. Retrieved Oct 31, 2019 from https://www. faceplusplus.com/attributes/
- Timnit Gebru, Jamie Morgenstern, Briana Vecchione, JenniferWortman Vaughan, HannaWallach, Hal Daumeé III, and Kate Crawford. 2018. Datasheets for datasets. arXiv preprint arXiv:1803.09010 (2018). · arXiv
- Yandong Guo, Lei Zhang, Yuxiao Hu, Xiaodong He, and Jianfeng Gao. 2016. Ms-celeb-1m: Challenge of recognizing one million celebrities in the real world. Electronic imaging 2016, 11 (2016), 1--6.
- Amy Hawkins. 2018. Beijing's Big Brother Tech Needs African Faces. Retrieved October 31, 2019 from https://foreignpolicy.com/2018/07/24/beijings-big-brothertech- needs-african-faces/
- Michael Hind, Sameep Mehta, Aleksandra Mojsilovic, Ravi Nair, Karthikeyan Natesan Ramamurthy, Alexandra Olteanu, and Kush R Varshney. 2018. Increasing Trust in AI Services through Supplier's Declarations of Conformity. arXiv preprint arXiv:1808.07261 (2018). · arXiv
- IBM. 2019. Release notes. Retrieved Oct 31, 2019 from https://cloud.ibm.com/ docs/services/visual-recognition?topic=visual-recognition-release-notes
- Kairos. 2019. Kairos Face Recognition Pricing Guide. Retrieved Oct 31, 2019 from https://www.kairos.com/pricing
- Steven Melendez. 2018. Despite a surge of tech activism, Clarifai plans to push further into government work. (2018).
- Margaret Mitchell, SimoneWu, Andrew Zaldivar, Parker Barnes, Lucy Vasserman, Ben Hutchinson, Elena Spitzer, Inioluwa Deborah Raji, and Timnit Gebru. 2018. Model Cards for Model Reporting. CoRR abs/1810.03993 (2018). arXiv:1810.03993 http://arxiv.org/abs/1810.03993 · arXiv
Cited by
Learning Transferable Visual Models From Natural Language Supervision
Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.00020 · Cited by 15,293
“…the FairFace benchmark dataset for gender, race and age classification of images by intersectional categories, accuracy on benchmarks offers only one approximation of algorithmic fairness, as Raji et al. (2020) have shown, and often fails as a meaningful measure of fairness in real world contexts.”
BackgroundAhmad, K., Maabreh, M., Ghaly, M., Khan, K., Qadir, J., & Al-Fuqaha, A. (2022) · Computer Science Review · doi:10.1016/j.cosrev.2021.100452 · Cited by 291
Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction
Shelby, R., Rismani, S., Henne, K., Moon, A., Rostamzadeh, N., Nicholas, P., Yilla-Akbari, N., Gallegos, J., Smart, A., Garcia, E., et al. (2023) · Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3600211.3604673 · Cited by 279
“, facial features, skin tone, or voice), such as computer vision [46, 176], natural language processing [113, 131, 173], and speech recognition systems [124, 143].”
BackgroundWho Audits the Auditors? Recommendations from a field scan of the algorithmic auditing ecosystem
Costanza-Chock, S., Raji, I. D., & Buolamwini, J. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533213 · Cited by 277
Self-citationTo chat or bot to chat: Ethical issues with using chatbots in mental health
Coghlan, S., Leins, K., Sheldrick, S., Cheong, M., Gooding, P., & D'Alfonso, S. (2023) · DIGITAL HEALTH · doi:10.1177/20552076231183542 · Cited by 215
Metadata
- Authors
- Raji, I. D., Gebru, T., Mitchell, M., Buolamwini, J., Lee, J., & Denton, R.
- Affiliations
- University of Toronto (CA)
- Google (United States) (US)
- Massachusetts Institute of Technology (US)
- Google Research, San Francisco, CA, USA
- Venue
- Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Publisher
- ACM
- Published
- 2020-02-07
- Pages
- 145-151
- ISSN
- 3065-8365 (ISSN-L)
- OpenAlex
- W3004542466
- Cited by
- 299
- References
- 18
- Citations by year
- 2026: 46 · 2025: 46 · 2024: 50 · 2023: 54 · 2022: 52 · 2021: 44 · 2020: 6
- FWCI
- 40.45
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Proceedings of the AAAI/ACM Conference on AI Ethics and Society (publishedVersion, public-domain)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Keywords
- algorithmic auditing, face recognition, biometric systems, algorithmic fairness, unintended harm, ethical considerations, sensitive data, algorithmic accountability
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing. https://app.cochat.ai/papers/10.1145/3375627.3375820
- The paper
- Raji, I. D., Gebru, T., Mitchell, M., Buolamwini, J., Lee, J., & Denton, R. (2020). Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing. Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society. https://doi.org/10.1145/3375627.3375820
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.