Everyday Algorithm Auditing: Understanding the Power of Everyday Users in Surfacing Harmful Algorithmic Behaviors
Open accessThe DOI matches its Crossref record; 55 of the 55 references it cites were found in a registry or an index.
Abstract
A growing body of literature has proposed formal approaches to audit algorithmic systems for biased and harmful behaviors. While formal auditing approaches have been greatly impactful, they often suffer major blindspots, with critical issues surfacing only in the context of everyday use once systems are deployed. Recent years have seen many cases in which everyday users of algorithmic systems detect and raise awareness about harmful behaviors that they encounter in the course of their everyday interactions with these systems. However, to date little academic attention has been granted to these bottom-up, user-driven auditing processes. In this paper, we propose and explore the concept of everyday algorithm auditing, a process in which users detect, understand, and interrogate problematic machine behaviors via their day-to-day interactions with algorithmic systems. We argue that everyday users are powerful in surfacing problematic machine behaviors that may elude detection via more centrally-organized forms of auditing, regardless of users' knowledge about the underlying algorithms. We analyze several real-world cases of everyday algorithm auditing, drawing lessons from these cases for the design of future platforms and tools that facilitate such auditing behaviors. Finally, we discuss work that lies ahead, toward bridging the gaps between formal auditing approaches and the organic auditing behaviors that emerge in everyday use of algorithmic systems.
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
“Furthermore, by providing a basis on which technology providers can be held accountable, audits incentivise investments in adequate risk management (Shen et al., 2021).”
From Auditing of AI: Legal, Ethical and Technical Approaches (2023) · cites this paper as Shen et al., 2021
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 [138]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Indeed, a recent line of HCI research has explored the idea of”
From Out of Context: Investigating the Bias and Fairness Concerns of “Artificial Intelligence as a Service” (2023) · cites this paper as [138]
The cited source says what the claim attributes to it.
In the citing paper
“Recent HCI work on user auditing [132], such as the 'WeAudit' tool [126], exemplifies this shift and demonstrates the possibility of designing for a more participatory AI audit process [2, 52, 132].”
From Towards AI Accountability Infrastructure: Gaps and Opportunities in AI Audit Tooling (2025) · cites this paper as [52]
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
Rethinking the Public Sphere: A Contribution to the Critique of Actually Existing Democracy
Fraser, N. (1990) · Social Text · doi:10.2307/466240 · Cited by 5,147
Algorithms of Oppression: How Search Engines Reinforce Racism
Noble, S. U. (2020) · doi:10.2307/j.ctt1pwt9w5 · Cited by 2,709
Raji, I. D., Smart, A., White, R. N., Mitchell, M., Gebru, T., Hutchinson, B., Smith-Loud, J., Theron, D., & Barnes, P. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3372873 · Cited by 2,245 · cited 1×
“In response to these concerns, scholars have proposed a number of approaches to audit algorithmic systems for biased, discriminatory, or otherwise harmful behaviors 1 — both internally, within the organizations responsible for developing and maintaining these systems, and externally [73, 76].”
BackgroundFairness and Abstraction in Sociotechnical Systems
Selbst, A. D., Boyd, D., Friedler, S. A., Venkatasubramanian, S., & Vertesi, J. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287598 · Cited by 1,631 · cited 2×
“Other harmful behaviors may only emerge when a system is used in unanticipated ways or in unanticipated contexts, perhaps due to changing norms and practices surrounding the use of a given algorithmic system over time [18, 82].”
BackgroundMeme-tracking and the dynamics of the news cycle
Leskovec, J., Backstrom, L., & Kleinberg, J. (2009) · Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/1557019.1557077 · Cited by 1,530 · cited 1×
“This is a particularly appealing possibility given that information in memes tends to propagate from mainstream news to social media [58].”
BackgroundPredicting protein structures with a multiplayer online game
Cooper, S., Khatib, F., Treuille, A., Barbero, J., Lee, J., Beenen, M., Leaver-Fay, A., Baker, D., Popović, Z., & players, F. (2010) · Nature · doi:10.1038/nature09304 · Cited by 1,368 · cited 1×
“Past work has explored collective sensemaking across a range of contexts, including citizen science [17], image labelling [96], knowledge mapping and curation (e.”
BackgroundFriedman, B., & Nissenbaum, H. (1996) · ACM Transactions on Information Systems · doi:10.1145/230538.230561 · Cited by 1,203
Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women *
Dastin, J. (2022) · Ethics of Data and Analytics · doi:10.1201/9781003278290-44 · Cited by 1,165
Algorithms as culture: Some tactics for the ethnography of algorithmic systems
Seaver, N. (2017) · Big Data & Society · doi:10.1177/2053951717738104 · Cited by 1,010 · cited 3×
“These factors make detecting certain harmful algorithmic behaviors difficult for auditors unless they are embedded within particular real-world social or cultural contexts [18, 49, 80, 101].”
BackgroundMemes in a Digital World: Reconciling with a Conceptual Troublemaker
Shifman, L. (2013) · Journal of Computer-Mediated Communication · doi:10.1111/jcc4.12013 · Cited by 898
Research works printed but not linked: 29
- Todd Bishop. 2015. Amazon changes its key formula for calculating product ratings and displaying reviews. GeekWire (June 2015). https://www.geekwire.com/2015/amazon-changes-its-influential-formula-for-calculating-product- ratings
- Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. 2020. Language (technology) is power: A critical survey of" bias" in nlp. arXiv preprint arXiv:2005.14050 (2020). · arXiv
- Rucksack Brian. 2019. Major changes to review scoring method on Booking com. https://hostelmanagement.com/f orums/major-changes-review-scoring-method-booking-com.html
- Joy Buolamwini. 2016. How I'm fighting bias in algorithms. November, TEDx Beacon Street)[Video File] (2016).
- Joy Buolamwini and Timnit Gebru. 2018. Gender shades: Intersectional accuracy disparities in commercial gender classification. In Proceedings of the 1st Conference on Fairness, Accountability and Transparency PMLR. 77-91.
- Kate Crawford and Trevor Paglen. 2021. Excavating AI: The politics of images in machine learning training sets. AI & Society (2021), 1-12.
- Richard Dawkins. 2016. The selfish gene. Oxford University Press.
- Michel De Certeau. 1984. The practice of everyday life, trans Steven Rendall. Berkeley: University of California Press.
- Nicholas Diakopoulos. 2014. Algorithmic accountability reporting: On the investigation of black boxes. The Tow Center for Digital Journalism (2014).
- Geoffrey A. Fowler. 2011. Judge dismisses suit against Yelp. The Wall Street Journal (Oct. 2011). https://www.wsj.co m/articles/SB10001424052970204505304577002170423750412
- Bryce Goodman and Seth Flaxman. 2016. EU regulations on algorithmic decision-making and a "right to explanation". In ICML Workshop on Human Interpretability in Machine Learning.
- Jessica Guynn. 2015. Google photos labeled black people 'gorillas'. USA Today (June 2015). https://www.usatoday.c om/story/tech/2015/07/01/google-apologizes-after-photos-identify-black-people-as-gorillas/29567465
- Jürgen Habermas. 1989. The structural transformation of the public sphere: An inquiry into a category of bourgeois society. MIT press.
- White House. 2016. Big Data: A Report on Algorithmic Systems. Opportunity, and Civil Rights (2016).
- Inkoo Kang. 2013. Businesses: 'Yelp is the thug of the Internet'. MuckRock (2013). https://www.muckrock.com/news/ archives/2013/jan/23/businesses-yelp-thug-of-the-internet.
- Inkoo Kang. 2013. Read nearly 700 FTC complaints regarding Yelp. MuckRock (Jan 2013). https://www.muckrock.c om/news/archives/2013/jan/23/businesses-yelp-thug-of-the-internet/
- Ronald B Mincy. 1993. The Urban Institute audit studies: Their research and policy context. In Clear and convincing evidence: Measurement of discrimination in America, M. Fix and R. J. Struyk (Eds.). Urban Institute Press, 165-86.
- Parmy Olson. 2018. The algorithm that helped google translate become sexist. Forbes (Feb. 2018). https://www.forbes .com/sites/parmyolson/2018/02/15/the-algorithm-that-helped-google-translate-become-sexist/?sh=6c1d0807daa2
- Andy Ozment. 2004. Bug auctions: Vulnerability markets reconsidered. In Third Workshop on the Economics of Information Security. 19-26.
- Charles C. Ragin and Howard S. Becker. 1992. What is a case?: Exploring the foundations of social inquiry. Cambridge University Press.
- Christian Sandvig, Kevin Hamilton, Karrie Karahalios, and Cedric Langbort. 2014. Auditing algorithms: Research methods for detecting discrimination on internet platforms. Data and Discrimination: Converting Critical Concerns into Productive Inquiry (2014).
- Satviki Sanjay. 2020. Twitter's AI Revealed to Have Racial Bias After a Viral Photo Experiment. Vice (Sep 2020). https://www.vice.com/en/article/93547v/twitter-ai-algorithm-has-racial-bias
- James C Scott. 1985. Weapons of the weak: Everyday forms of peasant resistance. Yale University Press.
- Nick Seaver. 2019. Knowing algorithms. Digital STS (2019), 412-422.
- Lucy A Suchman. 1987. Plans and situated actions: The problem of human-machine communication. Cambridge University Press.
- Harry Tucker. 2016. Australian Uber drivers say the company is manipulating their ratings to boost its fees. Businessinsider (May 2016). https://www.businessinsider.com.au/australian-uber-drivers-say-the-company-is- manipulating-their-ratings-to-boost-the-companys-fees-2016-5
- Colin Vandenhof and Edith Law. 2019. Contradict the Machine: A Hybrid Approach to Identifying Unknown Unknowns. In Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems. 2238-2240.
- Julia Velkova and Anne Kaun. 2019. Algorithmic resistance: Media practices and the politics of repair. Information, Communication & Society (2019), 1-18.
- Karl E Weick. 1995. Sensemaking in organizations. Vol. 3. Sage.
Cited by
Items 101–200 of 266
Panigutti, C., Yela, D. F., Porcaro, L., Bertrand, A., & Garrido, J. S. (2025) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3715275.3732052 · Cited by 14
FairFare : A Tool for Crowdsourcing Rideshare Data to Empower Labor Organizers
Calacci, D., Rao, V. N., Dalal, S., Di, C., Pua, K.-W., Schwartz, A., Spitzberg, D., & Monroy‐Hernández, A. (2026) · ACM Transactions on Computer-Human Interaction · doi:10.1145/3769680 · Cited by 14
“Later tools, such as the Shipt Calculator [13], attempted to reduce the human labor required for collective resistance by automating data collection and analysis.”
Making Algorithms Public: Reimagining Auditing from Matters of Fact to Matters of Concern
Geiger, R. S., Tandon, U., Gakhokidze, A., Song, L., & Irani, L. C. (2024) · eScholarship (California Digital Library) · Cited by 14
Pyle, C., Zhang, Z., Haimson, O. L., & Andalibi, N. (2024) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3637342 · Cited by 13
“Given the increasing work around algorithmic audits [89, 106], including user-led algorithmic audits [44, 111], as mechanisms for revealing harmful implications of existing algorithmic technologies, future work may investigate to what extent these kinds of auditing practices may leverage existing…”
BackgroundDeng, W. H., Yurrita, M., Díaz, M., Suh, J., Judd, N., Groves, L., Shen, H., Eslami, M., & Holstein, K. (2024) · Proceedings of the AAAI Conference on Human Computation and Crowdsourcing · doi:10.1609/hcomp.v12i1.31609 · Cited by 13
“This is largely due to the proven ability of non-experts with relatively low tech-savviness and AI literacy to uncover problematic AI behaviors that might be overlooked by small groups of AI researchers and developers (Shen et al. 2021; DeVos et al. 2022; Lam et al. 2022).”
BackgroundSelf-citationThe AI Surge: Transfiguring Audit Profession With AI Technology
Kinana, S. B. B., & Arabiat, O. S. (2025) · Emerald Group Publishing Limited eBooks · doi:10.1108/978-1-83608-756-420251004 · Cited by 13
Vuyyuru, S. R. (2025) · International Journal of Engineering Development and Research · doi:10.56975/ijedr.v13i2.301034 · Cited by 13
Learning AI Auditing: A Case Study of Teenagers Auditing a Generative AI Model
Morales‐Navarro, L., Gan, M. A., Yu, E., Vogelstein, L., Kafai, Y. B., & Metaxa, D. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757620 · Cited by 13
“Other literature observes that users may sometimes engage organically with auditing practices by evaluating algorithmic systems in their everyday lives in the absence of experts [13, 51].”
InfluentialAudit-as-code: a policy-as-code framework for continuous AI assurance
Muhammad, A. E., Yow, K., & Alsenan, S. A. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1759211 · Cited by 13
Zhong, H., Do, T., Jie, Y., Neuwirth, R. J., & Shen, H. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.04766 · Cited by 13
“This theme underscores the pivotal role of "everyday users" in algorithm auditing and value alignment [31, 32, 33, 34], illustrating how ordinary individuals can effectively identify and address biases in AI systems.”
Self-citationJohnson, N., Silva, E., Leon, H., Eslami, M., Schwanke, B., Dotan, R., & Heidari, H. (2025) · Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3715275.3732049 · Cited by 12
Self-citationIdentifying Algorithmic Decision Subjects' Needs for Meaningful Contestability
Yurrita, M., Verma, H., Balayn, A., Alfrink, K., Gadiraju, U., & Bozzon, A. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757415 · Cited by 13
“In a similar vein, end-user driven audits [29, 30, 59, 89] use the lived experiences of everyday users of algorithmic systems to uncover harmful algorithmic behaviors, which has, in turn, led to collective contestations (see Shen et al. [89] for a list of examples).”
Han, E. X., Zhang, A. Q., Zhu, H., Shen, H., Liang, P. P., & Hsieh, J. (2025) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3746059.3747710 · Cited by 12
“To expand the output spaces of text-to-image models and to reduce the burden on users to manually refine prompts, we investigate ways to automatically detect and expand their homogeneous dimensions , so that users can explore more varied design alternatives with less effort.”
Self-citationTaylor, J., Mire, J., DeVrio, A., Sap, M., Zhu, H., & Fox, S. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3816950 · Cited by 12
“At the application level, developers might imperceptibly perturb images and videos that users upload [109] to protect these works from being used to build GenAI models, similar to social media sites removing location metadata when users upload images [39, 54].”
Algorithmic Collective Action with Two Collectives
Karan, A., Vincent, N., Karahalios, K. G., & Sundaram, H. (2025) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3715275.3732098 · Cited by 11
“Understanding and Countering Impacts of Algorithms: A number of scholars in computing have sought to understand how people engage in forms of algorithmic resistance, using lenses such as folk theory [8, 9, 29] and auditing [11, 49].”
Xiao, Q., Chen, R., Xiao, J., Fu, T., Zhang, A. Q., Fan, X., Zhang, B., Lu, Z., & Shen, H. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3816914 · Cited by 12
“…CSCW literature has emphasized algorithmic interpretation as emergent, user-driven responses to algorithmic opacity (e.g., folk theories) [21, 38, 59], our study shifts the analytic focus to institutional actors and their role in intentionally constructing and instrumentalizing such algorithmic…”
Self-citationWhat Comes After Harm? Mapping Reparative Actions in AI Through Justice Frameworks
Xiao, S., Zou, H., Zhang, A. Q., Kumar, D., Shen, H., Hong, J., & Eslami, M. (2025) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1609/aies.v8i3.36754 · Cited by 10
Self-citationAllen, L. K., DeLiema, D., & Kendeou, P. (2025) · Educational Psychologist · doi:10.1080/00461520.2025.2588213 · Cited by 10
“By contrast, learners with stronger algorithmic awareness may avoid equating ranking with credibility, deliberately examine lower-ranked links, or even engage in public audits of algorithmic influence (Shen et al., 2021).”
InfluentialYee, K., Sebag, A. S., Redfield, O., Eck, M., Sheng, E., & Belli, L. S. (2023) · Workshop on Trustworthy Natural Language Processing (TrustNLP) · doi:10.18653/v1/2023.trustnlp-1.10 · Cited by 9
Reflective Design for Informal Participatory Algorithm Auditing: A Case Study with Emotion AI
Howell, N., Hartsoe, W. F., Amin, J., & Namani, V. (2024) · Nordic Conference on Human-Computer Interaction · doi:10.1145/3679318.3685411 · Cited by 9
“In some cases, the algorithm may be attempting to replicate a task with clear answers that most people can identify and agree on, such as image recognition of commonplace objects (e.g., [61, 128]).”
InfluentialGeiger, R. S., O’Sullivan, F., Wang, E., & Lo, J. (2025) · PLoS ONE · doi:10.1371/journal.pone.0318500 · Cited by 9
Hulstijn, J. (2023) · International Conference on Artificial Intelligence and Law · doi:10.1145/3594536.3595122 · Cited by 9
“A promising approach is called everyday algorithm auditing [64].”
BackgroundStudying Collaborative Interactive Machine Teaching in Image Classification
Mohammadzadeh, B., Françoise, J., Gouiffès, M., & Caramiaux, B. (2024) · ACM International Conference on Intelligent User Interfaces (IUI) · doi:10.1145/3640543.3645204 · Cited by 9
Exploring Subjectivity for more Human-Centric Assessment of Social Biases in Large Language ModelsPreprint
Aoyagui, P. A., Ferguson, S. A., & Kuzminykh, A. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2405.11048 · Cited by 9
“One potential avenue to be explored is leveraging human evaluation when end-users interact with a Large Language Model (LLM), similar to the one proposed by Shen et al. [43].”
Aggregated Individual Reporting for Post-Deployment EvaluationPreprint
Dai, J., Raji, I. D., Recht, B., & Chen, I. Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2506.18133 · Cited by 9
“We are also heavily influenced by calls to incorporate end-user expertise in algorithm evaluations (e.g., Shen et al. [2021], DeVos et al. [2022], Lam et al. [2022]) and to shift power towards the public (e.g., Kalluri [2020], Feffer et al. [2023]).”
InfluentialZhou, K., Gligorić, K., Cheng, M., Lam, M. S., Raman, V., Aminu, B., Woo, C., Brockman, M., Cha, H., & Jurafsky, D. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.67 · Cited by 8
Geiger, R. S., Tandon, U., Gakhokidze, A., Song, L., & Irani, L. C. (2023) · International journal of communication · Cited by 8
Delmonaco, D. (2023) · Deep Blue (University of Michigan) · doi:10.7302/8386 · Cited by 8
Ethics-based auditing of automated decision-making systems: considerations, challenges, na
Mökander, J. (2023) · Oxford University Research Archive (ORA) (University of Oxford) · doi:10.5287/ora-z5qakao0p · Cited by 12
System Cards for AI-Based Decision-Making for Public Policy
Gursoy, F., & Kakadiaris, I. A. (2022) · arXiv · doi:10.48550/arxiv.2203.04754 · Cited by 7
“On the other hand, some studies focus on particular themes such as nondiscrimination [30], [37], [38], data [37], [39], models [40], transparency [34], human rights [41], reputational concerns [42], audits by everyday users [43], and assurance [44], [45].”
BackgroundHollanek, T., Pi, Y., Fiorini, C., Vignali, V., Peters, D., & Drage, E. (2025) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3715275.3732078 · Cited by 7
De, A., & Cotter, K. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757440 · Cited by 7
“While activists often find creative ways to work with algorithms to increase the visibility of their causes, they also face challenges in understanding these algorithms and developing effective tactics [118, 125, 133].”
Hong, J., Ko, E.-Y., Kim, J., & Jang, J.-W. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757706 · Cited by 7
“However, central moderation may fail to reflect individual user variances in tolerance against problematic content [28], and personalized curation may also fall short of accurately reflecting users’ preferences [42, 53].”
Mapping the Smart City: Participatory approaches to XAI
Gupta, S. (2023) · Designing Interactive Systems Conference · doi:10.1145/3563703.3593063 · Cited by 7
“Human cognitive abilities [66], users’ explanatory needs [37, 62], users’ situated real-world experiences [16], users’ ability to collaborate and form counter publics [57] have been some of the guiding factors in advancing user-centered XAI research.”
BackgroundCo-Designing Islamic AI Ethics: Insights from the UK Muslim Community
Campbell-Esen, R., Mahoney, J., & Talhouk, R. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790413 · Cited by 7
“Additionally, sh¯ ur¯ a can be part of end-user auditing [78] that rather than solely focusing on harms [122] facilitates a multi-perspective approach to assessing the public-benefits (maslaha) of AI systems, especially those to be deployed in the public sector, through the lens of collective…”
Humane Autonomous Technology: Re-thinking Experience with and in Intelligent Systems
Rousi, R. A., von Koskull, C., & Roto, V. (2024) · Springer eBooks · doi:10.1007/978-3-031-66528-8 · Cited by 7
“That is one reason why deep neural networks are difficult to audit (Perez et al., 2022) and harm is often only discovered during everyday use (Shen, DeVos, Eslami, & Holstein, 2021).”
BackgroundThe Cultivated Practices of Text-to-Image GenerationPreprint
Oppenlaender, J. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2306.11393 · Cited by 7
“That is one reason why deep neural networks are difficult to audit (Perez et al., 2022) and harm is often only discovered during everyday use (Shen, DeVos, Eslami, & Holstein, 2021).”
BackgroundKrause, D. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5050631 · Cited by 7
Global Approaches to Auditing Artificial Intelligence: A Literature Review
(IPIE), I. P. O. T. I. E., Sylvester, P., Chun, W. H. K., Hannák, A., Mendoza, M., Broussard, M., Сергеевич, Е. Р., Nelson, A., Sandvig, C., Sesan, G., et al. (2024) · doi:10.61452/bwym7397 · Cited by 7
Wu, M., Zhao, Y., Han, S., Liu, M. X., & Shen, H. Y. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3816897 · Cited by 6
Self-citationPatterns, Models, and Challenges in Online Social Media: A Survey
Marco, N. D., Bonetti, A., Martino, E. D., Loru, E., Nudo, J., Pandolfo, M. E., Pecile, G., Sangiorgio, E., Scalco, I., Zollo, S., et al. (2026) · ACM Transactions on the Web · doi:10.1145/3796548 · Cited by 6
Bureaucratic Silences: What the Canadian AI Register Reveals, Omits, and Obscures
Das, D., Tessono, C., Ahmed, S. I., & Guha, S. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812336 · Cited by 8
“Both researchers and everyday users have used algorithmic audits to scrutinize AI systems [29, 69, 115], like recommendation and search algorithms [10, 102], computer vision-based processes (e.g., generative art [120], image captioning [135], facial recognition [22]), and language technologies…”
Zhang, R., & Xie, L. (2026) · International Journal of Human-Computer Studies · doi:10.1016/j.ijhcs.2026.103897 · Cited by 6
“Studies have found that people adopt AI-specific sensemaking strategies, such as exchanging tips and tricks with fellow users, experimenting with prompt engineering, or tracking system updates to anticipate behavioral changes (DeVos et al., 2022; Shen et al., 2021).”
Vipera: Towards systematic auditing of generative text-to-image models at scale
Huang, Y., Deng, W. H., Xiao, S., Eslami, M., Hong, J., & Perer, A. (2025) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3706599.3719757 · Cited by 5
“However, characterizing the relationships between objects using plain text can be challenging.”
Self-citationDesign(ing) Fictions for Collective Civic Reporting of Privacy Harms
Wu, Y., Agnew, W. S., Edwards, W. K., & Das, S. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3710961 · Cited by 6
“Future work could explore how a GoTCHa for privacy might look for other kinds of privacy harms, such as data breaches and workplace surveillance, or how concepts such as everyday algorithmic auditing [52] might be institutionalized.”
Explainable pedagogical alignment: auditing LLMs with learning theories
Figaredo, D. D., Vindel, J. L. F., & Puerta, L. G. (2025) · Education and Information Technologies · doi:10.1007/s10639-025-13857-2 · Cited by 5
PolicyPad: Collaborative Prototyping of LLM Policies
Feng, K. J. K., Kuo, T.-S., Chen, Q. Z., Cheong, I., Holstein, K., & Zhang, A. X. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791689 · Cited by 5
Self-citationVito, P. D. C. S., Fringi, E., Johnston, P., Bezerra, L. C. T., Aristodemou, M., Shahandashti, S. F., O’Hara, E., Whyte, L. F., Luo, L. K., Wong, M., et al. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791757 · Cited by 5
“…ones, could easily be categorised using the taxonomy by Weidinger et al. [62], which shows that participatory auditing can broaden the typical understanding of implications of AI applications, highlighting and confirming previous findings about the usefulness of participatory auditing [23, 55].”
Liao, J., Wei, Z., Yang, Z., Xu, X., Hui, P., He, C., & Zhou, M. (2025) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3713763 · Cited by 6
“In response to this opacity, content creators have increasingly engaged in a form of algorithm auditing—informal efforts to understand and navigate the platform’s algorithms [2, 13, 26, 27, 58, 82].”
Schmude, T., Yurrita, M., Alfrink, K., Goff, T. L., Tschiatschek, S., & Viard, T. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812309 · Cited by 5
“Examples of such mechanisms include data input control [56], decision revision requests [59], and various audit methodologies [29, 64, 80, 90].”
Digital Platforms and Algorithmic Erasure: What are the Implications?
Madland, C., & Ofosuhene, M. (2022) · The Open/Technology in Education Society and Scholarship Association Conference · doi:10.18357/otessac.2022.2.1.137 · Cited by 5
MIRAGE: Multi-model Interface for Reviewing and Auditing Generative Text-to-Image AIPreprint
Maldaner, M. K., Deng, W. H., Hong, J., Holstein, K., & Eslami, M. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.19252 · Cited by 5
“This reluctance limits the ability of companies to gather valuable feedback from everyday users, who are social actors engaged in the daily use of algorithm auditing systems (Shen et al. 2021).”
Self-citationHuang, Y., Deng, W. H., Xiao, S., Eslami, M., Hong, J., Narechania, A. A., & Perer, A. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791942 · Cited by 4
“Prior works on AI auditing have explored various types of guidance ranging from community, algorithmic, to expert [17, 64].”
InfluentialSelf-citationSchmude, T., Koesten, L. M., Möller, T. B., & Tschiatschek, S. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2650562 · Cited by 4
Eschebach, T., Banović, N., & McDonald, A. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757496 · Cited by 4
“Though this practice is common among YouTubers [15, 68, 134], P10 used it as a way to conduct “everyday algorithm auditing” [121] to see if their perceptions of the algorithm were true.”
Making Smart Cities Explainable: What XAI Can Learn from the “Ghost Map”
Gupta, S., & Loukissas, Y. A. (2023) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3544549.3585847 · Cited by 4
“Users may need guidance from experts or algorithms to better audit algorithmic systems [48].”
BackgroundWeAudit: Scaffolding User Auditors and AI Practitioners in Auditing Generative AIPreprint
Deng, W. H., Wang, C., Han, H. Z., Hong, J., Holstein, K., & Eslami, M. (2025) · Qeios · doi:10.32388/b1lxnv · Cited by 4
Self-citationTowards Meaningful Transparency in Civic AI SystemsPreprint
Murray-Rust, D., Alfrink, K., & Zaga, C. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.07889 · Cited by 4
“One study argues that “everyday users of algorithmic systems are often best poised to detect and raise awareness about harmful behaviors” they encounter (Shen et al., 2021; Van Nuenen et al., 2022).”
Han, E. X., Zhang, A. Q., Zhu, H., Shen, H., Liang, P. P., & Hsieh, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.13392 · Cited by 4
“To expand the output spaces of text-to-image models and to reduce the burden on users to manually refine prompts, we investigate ways to automatically detect and expand their homogeneous dimensions , so that users can explore more varied design alternatives with less effort.”
Self-citationPromptimizer: User-Led Prompt Optimization for Personal Content Classification
Wang, L., Yurechko, K., & Zhang, A. X. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790923 · Cited by 3
“We consider initializing and iterating on personal content classifiers as two distinct scenarios, where how users communicate their preferences can vary greatly.”
Surveillance, Spacing, Screaming and Scabbing: How Digital Technology Facilitates Union Busting
Reiber, F., Kim, N., McDonald, A., & Calacci, D. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791507 · Cited by 3
“Our third criterion followed from other HCI research that has adopted a case study approach [153], namely choosing cases that were (a) accessible to us via multiple data sources and (b) spanned dimensions previously identified as important when studying labor struggles, including precarity, union…”
Push and Pushback in Contesting AI: Demands for and Resistance to Accountability
Pi, Y., Lichner, L., Lee, J. W., Xiao, S., Zhang, R., & Singh, J. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812323 · Cited by 3
“One common way individuals and communities from below sometimes push back against algorithmic systems is through algorithmic resistance – inventive, often tactical strategies that reactively subvert, evade, or mitigate algorithmic power [15, 26, 99, 120].”
InfluentialWang, S., Sun, X., Ni, S., Wu, M., & Hu, K. (2026) · Scientific Reports · doi:10.1038/s41598-026-61031-6 · Cited by 3
QA-TOOLBOX: Conversational Question-Answering for process task guidance in manufacturingPreprint
Manuvinakurike, R., Watkins, E. A., Savur, C., Rhodes, A., Biswas, S., Mejia, G. G., Beckwith, R., Sahay, S., Raffa, G., & Nachman, L. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2412.02638 · Cited by 3
From Seed to Harvest: Augmenting Human Creativity with AI for Red-teaming Text-to-Image ModelsPreprint
Quaye, J. A., Rastogi, C., Parrish, A., Inel, O., Kahng, M., Aroyo, L. M., & Reddi, V. J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2507.17922 · Cited by 3
“…or academic researchers often lacks coverage and diversity in failures identified, compared to the issues surfaced by the crowd [13, 20, 53], and small-group approaches may fail to find problematic model behaviors because participants lack the cultural background to identify certain…”
Choi, J., Choi, D., Jeong, S., Hong, H., & Seering, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.12268 · Cited by 3
“While CAI is often positioned as progressing toward greater inclusivity in society, recent research has nevertheless underscored the persistence of biased perspectives embedded within such systems [29, 30, 40, 77].”
Zhang, Z., Guo, B., & Li, T. (2024) · arXiv · doi:10.48550/arxiv.2411.01344 · Cited by 2
“Another promising approach is user auditing, such as real-time auditing that allows users to review actions when they use LM agents to complete a task [59], or post-hoc auditing where users verify some past examples or simulated examples at scale [28].”
'Take Everything With a Pinch of Salt': Learning to Navigate Fitness on Social Media
Dennehy, D. P., Murphy, S., Foley, S., McCarthy, J., & Morrissey, K. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3711080 · Cited by 2
“Although there may be benefits to engaging with fitness content on social media for women, this engagement is impacted by the actions of algorithms and recommender systems [85].”
Le$bean or lesbian? A survey of marginalised users' motivations for obfuscation on TikTok
Ungless, E. L., Markl, N., & Roß, B. (2025) · Behaviour and Information Technology · doi:10.1080/0144929x.2025.2553152 · Cited by 2
“Such folk theories often develop from ‘everyday algorithm auditing’: the discovery of problematic behaviour through everyday interactions with a system (Brown 2021; Karizat et al. 2021; Shen et al. 2021; Simpson and Semaan 2021).”
Understanding, Challenging, and Demystifying Perceptions of Gig Worker Vulnerabilities
de Jong, S., Hsieh, J., Kuo, T.-S., Jacobsen, R. M., van Berkel, N., & Zhu, H. (2026) · Annual Symposium on Human-Computer Interaction for Work (CHIWORK) · doi:10.1145/3808045.3808071 · Cited by 2
Sandvig, C., Aidinoff, M., Nelson, A., Vaccaro, K., Karahalios, K. G., Bhandari, E., Biddle, E. R., Armstrong, L., Matias, J. N., Eslami, M., et al. (2026) · The MIT Press eBooks · doi:10.7551/mitpress/15997.001.0001 · Cited by 2
Self-citationNoh, D. J., Fields, D. A., Kafai, Y. B., & Metaxa, D. (2026) · Proceedings. · doi:10.22318/icls2026.716535 · Cited by 2
Upadhyay, S., Pruss, D., DeVrio, A., Gajos, K. Z., & Karusala, N. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791923 · Cited by 2
“…by a variety of different stakeholders, ranging from more standardized and centralized approaches typical in government or industry [34, 82, 94, 101] to more participatory ones often involving community groups [48, 60], labor unions [26, 36, 59, 83, 90], and end users [27, 38, 39, 62, 64, 89].”
Informal Embodied Auditing: Exploring Facial Emotion AI (FEAI) through Community Workshops
Li, X., Riggs, A. T., Dai, Z., Farmer, C. B., Morrison, K. G., & Howell, N. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790863 · Cited by 2
Whose Knowledge Counts? Co-Designing Community-Centered AI Auditing Tools with Educators in Hawai'i
Zhao, D., Cha, H., Ryan, M. J., Wang, A., Baker-Ramos, R., Helekahi-Kaiwi, E.-B., Diego, R., Hester, J. D., & Yang, D. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790958 · Cited by 2
2. Revisiting Transparency Efforts in Algorithmic Regimes
Eslami, M., & Heuer, H. (2024) · Amsterdam University Press eBooks · doi:10.1515/9789048556908-002 · Cited by 2
Self-citationWeb-Based Complaints Service Information System at Dewantara District Office
Amrina, A., Aslam, A., Islami, K., Sutapa, I. K., Kurniawansyah, K., Abdullah, D., & Ardian, Z. (2024) · International Journal of Engineering Science and Information Technology · doi:10.52088/ijesty.v4i1.490 · Cited by 2
“The definition of a system is that a system can be said to be a collection of functioning networks consisting of different elements connected to achieve a specific goal [17] [18].”
BackgroundLu, Y., Xie, Z., Chiu, D. K. W., & Kafeza, E. (2025) · Systems · doi:10.3390/systems13110975 · Cited by 2
Compass: Continuously Aligning Social Media Feeds via In-Situ Reflections
Barua, A., Wang, L., & Zhang, A. X. (2026) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3830398.3830671 · Cited by 2
“These interactions fit naturally into how casual social media users consume recommendation feeds by scanning and informally evaluating content as they browse [68].”
Three Essays on HRM Algorithms: Where Do We Go from Here?
Cheng, M. (2024) · MacSphere (McMaster University) · doi:10.71548/20919 · Cited by 2
A User-driven Design Framework for RobotaxiPreprint
Deng, Y., & He, C. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.19107 · Cited by 2
“AI [35, 68, 75].”
Kilhoffer, Z. (2025) · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · Cited by 2
When Your Reviewer is an LLM: Biases, Divergence, and Prompt Injection Risks in Peer Review
Zhu, C., Xiong, J., Ma, R., Lu, Z., Liu, Y., & Li, L. (2025) · arXiv · doi:10.48550/arxiv.2509.09912 · Cited by 1
“This approach embodies the principles of algorithmic auditing, empowering human reviewers to surface harmful behaviors in their context of practice [28, 55].”
Evaluating Text-to-Image Platforms' Content Moderation During the 2024 US Presidential Election
Greene, K. T., Norton, S. T., & Shapiro, J. N. (2025) · Journal of Quantitative Description Digital Media · doi:10.51685/jqd.2025.021 · Cited by 1
End-user Algorithmic Auditing for Music Discoverability: A Research Roadmap
Porcaro, L., Gómez, E., & Catarci, T. (2024) · IRIS Research product catalog (Sapienza University of Rome) · Cited by 1
Porcaro, L., Mirabella, V., Gómez, E., & Catarci, T. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791406 · Cited by 1
Human-Centered and Participatory AI Auditing
Deng, W. H., Holstein, K., & Eslami, M. (2026) · Springer eBooks · doi:10.1007/978-981-97-8440-0_102-1 · Cited by 1
Self-citationPeng, C., Chai, M., Mo, G., Raman, N. J., Tang, N., Pagdon, S., Swarbrick, M. A., Jones, N., Fang, F., & Shen, H. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790339 · Cited by 1
“Clear reflective procedures such as acknowledging errors, explaining why a suggestion worked or failed, noticing conflicts with peer knowledge, correcting courses, continuing to build expertise, and re-centering the relationship can turn break-downs into teachable moments [88].”
Self-citationMohammadzadeh, B., Françoise, J., Gouiffès, M., & Caramiaux, B. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790784 · Cited by 1
“For example, Twitter users tested hypotheses about racial bias in the plat-form’s image-cropping algorithm by altering features such as background color or clothing [17], and Booking.com users adjusted input ratings to uncover patterns of score inflation [39].”
InfluentialStandardization of AI Ethics: Stakeholders, Values and Profit
Corrêa, N. K., & Mönig, J. M. (2026) · Technosophy · doi:10.14361/9783839478431 · Cited by 1
Légitime, S., & Suresh, H. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3817010 · Cited by 1
“As a result, many researchers and communities have advanced strategies to pressure centralized platforms and regain some agency within the social medium by performing user audits [116], using data strikes and conscious data contributions [129, 130], and engaging in “algorithmic gossip” [33, 74, 92].”
Human-Centered and Participatory AI Auditing
Deng, W. H., Holstein, K., & Eslami, M. (2026) · Springer eBooks · doi:10.1007/978-981-95-3658-0_102 · Cited by 1
Self-citationBeyond Bias Detection: Community Auditors and Normative Reasoning in AI Oversight CSCW006
Jackson, C., Ahmad, T., Raj, S. D., & Wu, N. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3788042 · Cited by 1
Exposure to news and political content on TikTok: A Case Study Around the 2024 US Election
Ge, C., Gautam, N., Vajjhala, S., Ng, C., Hasell, A., & Tomkins, S. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812207 · Cited by 1
“We also acknowledge limitations of sock puppets discussed by others [5, 11, 27, 34, 52].”
Cha, H. (2026) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1609/aies.v9i1.43167 · Cited by 1
“Overall, these auditing methods focus on experiential harms, or harms that can be surfaced through self-reported experiences of negative affect [90], from the users themselves.”
No More Plaudits for Algorithm Audits: A Provocation
DeVrio, A., & Widder, D. G. (2026) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1609/aies.v9i2.43185 · Cited by 1
Register, Y., & Spiro, E. S. (2022) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v16i1.19337 · Cited by 1
“Because some of the most egregious harms tend to impact marginalized people specifically ((Shen et al. 2021; Oliva, Antonialli, and Gomes 2021; Alkhatib 2021; Are 2020; Haimson et al. 2021), we particularly consider the voices of marginalized individuals, the kinds of harms they identify and…”
BackgroundOptimizing Bug Bounty Programs for Efficient Malware-Related Vulnerability Discovery
Yulianto, S., Soewito, B., Gaol, F. L., & Kurniawan, A. (2024) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2024.0150430 · Cited by 1
“Bug bounty programs offer monetary and reputational rewards to hackers who share vulnerabilities with legitimate organizations, thereby deterring them from selling or exploiting these vulnerabilities [6].”
BackgroundRondina, M., Vetrò, A., Coppola, R., Regragui, O., Fabris, A., Silvello, G., Susto, G. A., & Martin, J. C. D. (2025) · Frontiers of artificial intelligence, ethics and multidisciplinary applications. · doi:10.1007/978-981-96-7945-4_15 · Cited by 1
Metadata
- Authors
- Shen, H., DeVos, A., Eslami, M., & Holstein, K.
- Affiliations
- Carnegie Mellon University (US)
- Carnegie Mellon University, Pittsburgh, PA, USA
- Venue
- Proceedings of the ACM on Human-Computer Interaction
- Published
- 2021-10-13
- Volume
- 5
- Issue
- CSCW2
- Pages
- 1-29
- ISSN
- 2573-0142 (ISSN-L)
- OpenAlex
- W3160887439
- Cited by
- 171
- References
- 49
- Citations by year
- 2026: 43 · 2025: 49 · 2024: 28 · 2023: 31 · 2022: 19 · 2021: 1
- FWCI
- 24.69
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- Proceedings of the ACM on Human-Computer Interaction (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Funders
- National Science Foundation (IIS-2040942)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
- Keywords
- algorithmic auditing, algorithmic bias, platform design
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
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.