Datasheets for datasets
Open accessThe DOI matches its Crossref record; 36 of the 37 references it cites were found in a registry or an index.
Abstract
Documentation to facilitate communication between dataset creators and consumers.
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
“Paradata needs may include the preservation of novel forms of information, such as datasheets explaining training datasets or model cards documenting an AI tool, and of these documents' relation to archival collections processed using them [ 26 ].”
From Positioning Paradata: A Conceptual Frame for AI Processual Documentation in Archives and Recordkeeping Contexts (2023) · cites this paper as [26]
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
“Significant critiques of LLMs have been raised, including concerns about bias and fairness, misinformation, opaqueness due to a lack of interpretability and explainability, and environmental impact (Bender et al., 2021; Gebru et al., 2021; Strubell et al., 2019).”
From Using ChatGPT and Other Large Language Model (LLM) Applications for Academic Paper Assignments (2023) · cites this paper as Gebru 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
“If ML driven algorithms are considered as epistemic resources that help us to improve understanding and/or make better predictions, then the statistical nature of ML de facto prevents groups underrepresented in the data from contributing to this epistemic resource on an equal ground to dominant groups (Gebru et al., 2021)”
From Algorithmic profiling as a source of hermeneutical injustice (2025) · cites this paper as Gebru 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
“The authors found that providing examples of what the requirements can look like helps to guide practitioners in clarifying their own rules.”
From A Survey of Data Quality Requirements That Matter in ML Development Pipelines (2023) · cites this paper as [24]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“I show how archival theories have been animated in documenting data in the context of AI and ML, including through documentation frameworks like Datasheets for Datasets”
From Can AI provenance inform archival provenance? (2026) · cites this paper as Gebru et al. 2021
12 more claims about this paper
Create a free account to see every claim other papers make about this one.
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
Advances in neural information processing systems 7
Cios, K., & Shields, M. (1997) · Neurocomputing · doi:10.1016/s0925-2312(97)90018-7 · Cited by 22,059
Labeled Faces in the Wild: A Database forStudying Face Recognition in Unconstrained Environments
Huang, G. B., Mattar, M., Berg, T. L., & Miller, E. G. (2008) · HAL · Cited by 4,822
Minimum information about a microarray experiment (MIAME)-toward standards for microarray data
Brazma, A., Hingamp, P., Quackenbush, J., Sherlock, G., Spellman, P., Stoeckert, C., Aach, J., Ansorge, W., Ball, C. A., Causton, H. C., et al. (2001) · Nature Genetics · doi:10.1038/ng1201-365 · Cited by 4,109
(2018) · Cited by 3,510
Self-citationA sentimental education: sentiment analysis using subjectivity summarization based on minimum cuts
Pang, B., & Lee, L. (2004) · Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics - ACL '04 · doi:10.3115/1218955.1218990 · Cited by 3,301
Semantics derived automatically from language corpora contain human-like biases
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017) · Science · doi:10.1126/science.aal4230 · Cited by 2,887
Model Cards for Model Reporting
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287596 · Cited by 2,681 · cited 1×
“Researchers at Google published follow-up work on model cards that document machine learning models [20] and released a data card (a lightweight version of a datasheet) along with the Open Images dataset [17].”
BackgroundSelf-citationMan is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
Bolukbasi, T., Chang, K.-W., Zou, J., Saligrama, V., & Kalai, A. (2016) · arXiv · doi:10.48550/arxiv.1607.06520 · Cited by 1,615 · cited 2×
“Of particular concern are recent examples showing that machine learning models can reproduce or amplify unwanted societal biases reflected in training data [4, 5, 9].”
BackgroundThe Effects of Automobile Safety Regulation
Peltzman, S. (1975) · Journal of Political Economy · doi:10.1086/260352 · Cited by 1,531
The Recent Past and Near Future of Risk and/or Need Assessment
Andrews, D. A., Bonta, J., & Wormith, J. S. (2006) · Crime & Delinquency · doi:10.1177/0011128705281756 · Cited by 1,507 · cited 1×
“Domains such as criminal justice (Garvie et al., 2016; Systems, 2017; Andrews et al., 2006), hiring and employment (Mann & O’Neil, 2016), critical infrastructure (O’Connor, 2017; Chui, 2017), and finance (Lin, 2012) all increasingly depend on machine learning methods.”
Background
Research works printed but not linked: 7
- Chui, G. Project will use AI to prevent or minimize electric grid failures, 2017.
- Garvie, C., Bedoya, A., and Frankle, J. The Perpetual Line-Up: Unregulated Police Face Recognition in America. Georgetown Law, Center on Privacy & Technology, Washington, D.C., 2016.
- Holstein, K., Vaughan, J., Daumé, H, Dudík, M., and Wallach, H. Improving fairness in machine learning systems: What do industry practitioners need? In Proceedings of 2019 ACM CHI Conf. on Human Factors in Computing Systems.
- Krasin, I. et al. OpenImages: A public dataset for large-scale multi-label and multi-class image classification, 2017.
- O'Connor, M. How AI Could Smarten Up Our Water System, 2017.
- Doha Supply Systems. Facial Recognition, 2017.
- Yagcioglu, S., Erdem, A., Erdem, E., and Ikizler-Cinbis, N. RecipeQA: A challenge dataset for multimodal comprehension of cooking recipes. In Proceedings of the 2018 Conf. on Empirical Methods in Natural Language Processing.
Cited by
Items 1,001–1,039 of 1,039
TAPS Responsibility matrix: a tool for responsible data science by design
Urovi, V., Çelebi, R., Sun, C., Kumar, P., Rieswijk, L., Erard, M., Yılmaz, A., Moodley, K., & Dumontier, M. (2024) · Journal of Responsible Innovation · doi:10.1080/23299460.2024.2414530 · Cited by 8
Nakayama, L. F., Ribeiro, L. Z., Malerbi, F. K., & Regatieri, C. V. (2025) · International Journal of Retina and Vitreous · doi:10.1186/s40942-024-00596-8 · Cited by 9
“Data is the cornerstone upon which AI advancements are built and the description of the included data for algorithm training and of datasets are critical for improving reproducibility and identify possible biases in AI development [9, 10].”
On the comparability between studies in predictive ecotoxicology
Schür, C., Schirmer, K., & Baity‐Jesi, M. (2025) · Computational Toxicology · doi:10.1016/j.comtox.2025.100367 · Cited by 7
An Overview of Fairness in Data – Illuminating the Bias in Data Pipeline
Kumar, B. S., Chandrabose, A., & Chakravarthi, B. R. (2021) · Cited by 17
Digital Ethics as Translational Ethics
Danks, D. J. (2021) · Advances in human and social aspects of technology book series · doi:10.4018/978-1-7998-8467-5.ch001 · Cited by 6
“Datasheets (Gebru et al., 2018) and model cards (Mitchell et al., 2019) are two process innovations that aim to reduce or eliminate problems that can arise from this kind of distributed work.”
BackgroundBrate, R., Nesterov, A., Vogelmann, V., van Ossenbruggen, J., Laura Hollink, A., & van Erp, M. (2021) · International Conference on Knowledge Capture · doi:10.1145/3460210.3493553 · Cited by 6
“The code and resources used to generate ConConCor, general documentation and the corpus’ accompanying datasheet (based on [7]) are available at https://github.”
MethodHäußler, H. (2021) · Libri · doi:10.1515/libri-2021-0095 · Cited by 10
“Another suggestion is disclosure of choices and assumptions during the technical design process that allow to judge the context (Gebru et al. 2021; Kraemer, van Over-veld, and Peterson 2011; Steen 2015).”
BackgroundObject classification on video data of meteors and meteor-like phenomena: algorithm and data
Sennlaub, R., Hofmann, M. R., Hankey, M., Ennes, M., Müller, T., Kroll, P. B., & Mäder, P. (2022) · Monthly Notices of the Royal Astronomical Society · doi:10.1093/mnras/stac1948 · Cited by 6
“A datasheet Gebru et al. (2018) summarizing NightSkyUCP along with detailed documentation can be found in the Appendix A.”
MethodNetwork cards: concise, readable summaries of network data
Bagrow, J. P., & Ahn, Y. (2022) · Applied Network Science · doi:10.1007/s41109-022-00514-7 · Cited by 5
“Network cards are also complementary to other efforts such as more detailed “datasheets” [7].”
InfluentialRevising human-systems engineering principles for embedded AI applications
Cummings, M. L. (2023) · Frontiers in Neuroergonomics · doi:10.3389/fnrgo.2023.1102165 · Cited by 8
“In terms of data curation, it is well-established that bias can be inadvertently introduced into an AI system due to underlying data sample selection bias (e.g., Samimi et al., 2010; Gianfrancesco et al., 2018).”
BackgroundWang, P., Xie, Y., Li, X., & Li, Q. (2023) · Journal of Librarianship and Information Science · doi:10.1177/09610006231154529 · Cited by 7
“There is evidence that providing adequate information about the data creation process and data content is the key for users to understand data (Gebru et al., 2021).”
Assessing the Impact of Cognitive Biases in AI Project Development
Bernault, C., Juan, S., Delmas, A., André, J., Rodier, M., & Kaadoud, I. C. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-35891-3_24 · Cited by 11
Pipeline Design for Data Preparation for Social Media Analysis
Bono, C. A., Cappiello, C., Pernici, B., Ramalli, E., & Vitali, M. (2023) · Journal of Data and Information Quality · doi:10.1145/3597305 · Cited by 9
Towards Labor Transparency in Situated Computational Systems Impact Research
Jing, F. S., Berger, S., & Sandoval, . C. B. (2023) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3593013.3594060 · Cited by 7
Ethics consideration sections in natural language processing papers
Benotti, L., & Blackburn, P. R. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.299 · Cited by 11
“…data usage (Drugan and Babych, 2010; Couillault et al., 2014; Mieskes, 2017; Bender and Friedman, 2018; Kann et al., 2019; Rogers et al., 2021; Gebru et al., 2021), crowdsourcing (Sny-der, 2010; Bederson and Quinn, 2011; Fort et al., 2011; Callison-Burch, 2014; Fort et al., 2014; Hara et al.,…”
BackgroundDeveloping responsible AI practices at the Smithsonian Institution
Dikow, R. B., DiPietro, C., Trizna, M. G., BredenbeckCorp, H., Bursell, M. G., Ekwealor, J. T. B., Hodel, R. G. J., Lopez, N., Mattingly, W., Munro, J., et al. (2023) · Research Ideas and Outcomes · doi:10.3897/rio.9.e113334 · Cited by 11
“Data statements (Bender and Friedman 2018, McMillan-Major et al. 2023), model cards (Mitchell et al. 2019), and datasheets (Gebru et al. 2021) are all focused on better documentation of data and methods, making any work resulting from their use more transparent and reproducible.”
BackgroundDesigning Fiduciary Artificial Intelligence
Benthall, S., & Shekman, D. (2023) · ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO) · doi:10.1145/3617694.3623230 · Cited by 25
“…frameworks have been proposed for guiding due diligence and accountability for machine learning based on the data set they use for training [47], conformity with testing standards [12], use of the model only for its intended purposes and with testing for fairness on varying demographic…”
MethodLiesenfeld, A., Lopez, A., & Dingemanse, M. (2023) · Annual Meeting of the Special Interest Group on Discourse and Dialogue · doi:10.18653/v1/2023.sigdial-1.45 · Cited by 9
Aguilar, N., Landau, A., Mathiyazhagan, S., Auyeung, A., Dillard, S., & Patton, D. U. (2023) · Hawaii International Conference on System Sciences (HICSS) · doi:10.24251/hicss.2023.090 · Cited by 10
Perdana, A., & Tan, S. L. (2024) · Springer eBooks · doi:10.1007/978-3-031-46209-2_12 · Cited by 10
Towards International Standards for the Evaluation of Artificial Intelligence for Health
Wenzel, M. A., & Wiegand, T. (2019) · ITU Kaleidoscope Academic Conference (ITU K) · doi:10.23919/ituk48006.2019.8996131 · Cited by 11
Updating the DCC Curation Lifecycle Model
Choudhury, S., Huang, C., & Palmer, C. L. (2020) · International Journal of Digital Curation · doi:10.2218/ijdc.v15i1.721 · Cited by 6
Towards evaluating and eliciting high-quality documentation for intelligent systems
Piorkowski, D., González, D., Richards, J., & Houde, S. (2020) · arXiv · doi:10.48550/arxiv.2011.08774 · Cited by 5
On the State of Reporting in Crowdsourcing Experiments and a Checklist to Aid Current Practices
Ramírez, J., Sayin, B., Baez, M., Casati, F., Cernuzzi, L., Benatallah, B., & Demartini, G. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3479531 · Cited by 4
“Examples of this can be found in evidence-based medicine where systematic reviews follow a strict elaboration protocol [71], or, more recently, in the machine learning community where authors are encouraged to follow pre-established checklists or datasheets to communicate their models and datasets effectively [2, 6, 2…”
BackgroundNLPositionality: Characterizing Design Biases of Datasets and Models
Santy, S., Liang, J. T., Bras, R. L., Reinecke, K., & Sap, M. (2023) · arXiv · doi:10.48550/arxiv.2306.01943 · Cited by 31
“The challenge is that design biases arise from the myriad of design choices made; in the context of creating datasets and models, only some of these choices may be documented (e.g., through model cards and data sheets; Bender and Friedman, 2018; Mitchell et al., 2019; Gebru et al., 2021).”
Background(2025) · doi:10.1145/3715275.3732087 · Cited by 4
Conversational AI Systems for Social Good: Opportunities and Challenges
Qi, P., Huang, J., Wu, Y., He, X., & Zhou, B. (2021) · arXiv · doi:10.48550/arxiv.2105.06457 · Cited by 3
“We urge researchers to consider using guidelines like Datasheets (Gebru et al. 2018) to help calibrate the collection and use of data, and when possible, involve members of communities the system is intended for in the process.”
MethodCross-Partisan Interactions on Twitter
Çetinkaya, Y. M., Ghafouri, V., Suarez-Tangil, G., Such, J., & Elmas, T. (2025) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v19i1.35819 · Cited by 16
E-GEO: A Testbed for Generative Engine Optimization in E-Commerce
Bagga, P. S., Farias, V. F., Korkotashvili, T., Peng, T., & Wu, Y. (2025) · arXiv · doi:10.48550/arxiv.2511.20867 · Cited by 1
Can AI provenance inform archival provenance?
Corry, F. (2026) · Cambridge Forum on AI: Culture and Society · doi:10.1017/cfc.2026.10025
“In AI/ML conversations, provenance typically refers to “data provenance,” or documentation about the “creation and use of datasets” for training computational models (Gebru et al. 2021, 86).”
InfluentialOutlining Traceability: A Principle for Operationalizing Accountability in Computing Systems
Kroll, J. A. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445937 · Cited by 2
“These include data sheets [101], fact sheets [10], data statements [28], data nutrition labels [123], model cards [171], and other standardized disclosure formats.”
BackgroundUnaccountable Delegation, Fading Skills: Mapping the Risks of Workplace AI Agents
La Malfa, G., Meegahapola, L., Bogucka, E., Zhang, J. M., Luck, M., Black, E., & Quercia, D. (2026) · arXiv · doi:10.48550/arxiv.2608.08601
“This task is routine in impact assessments, model cards, or risk registers (Mitchell et al. 2019; Gebru et al. 2021; Moss et al. 2021; Raji et al. 2020).”
Disclosure or Marketing? Analyzing the Efficacy of Vendor Self-reports for Vetting Public-sector AI
Kuehnert, B., Johnson, N., Dotan, R., & Heidari, H. (2026) · Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3805689.3812220
“Early work on documentation in machine learning such as Model Cards for Model Reporting [51] and Datasheets for Datasets [30] argue that structured documentation could uncover design decisions laden with implicit values, encourage responsible development practices, and provide stakeholders with…”
Algorithmic monoculture and social welfare
(2023) · The Societal Impacts of Algorithmic Decision-Making · doi:10.1145/3603195.3603202
Dead Science Walking: Publication Bias and the AI Scientist Pipeline
Chauhan, K. (2026) · arXiv · doi:10.48550/arxiv.2606.04220
“This follows the spirit of model cards, datasheets for datasets, data statements, data nutrition labels, and broader documentation standards for machine-learning artifacts (Mitchell et al., 2019; Gebru et al., 2021; Bender & Friedman, 2018; Hol-land et al., 2018; Raji et al., 2020).”
Designing Human-Automation Redundancy: Overreliance, Automation Shirking, and the Operator’s Dilemma
Cohen, D., Roth, Y., Rieskamp, J., & Schöbel, M. (2025) · doi:10.2139/ssrn.5098488
The Pile: An 800GB Dataset of Diverse Text for Language Modeling
Gao, L., Biderman, S., Black, S., Golding, L., Hoppe, T., Foster, C., Phang, J., He, H., Thite, A., Nabeshima, N., et al. (2021) · arXiv · doi:10.48550/arxiv.2101.00027 · Cited by 492
Large image datasets: A pyrrhic win for computer vision?
Prabhu, V. U., & Birhane, A. (2020) · arXiv · doi:10.48550/arxiv.2006.16923 · Cited by 91
“Much along the lines of model cards [55] and datasheet for datasets [35], we propose dissemination of dataset audit cards.”
MethodTowards Standardization of Data Licenses: The Montreal Data License
Benjamin, M., Gagnon, P., Rostamzadeh, N., Pal, C., Bengio, Y., & Shee, A. (2019) · arXiv · doi:10.48550/arxiv.1903.12262 · Cited by 26
“Conversely, recent efforts in standardizing the presentation of metadata for ML models are highly relevant, and contribute to fostering transparency in the fields of ML and AI (Gebru et al., 2018; Mitchell et al., 2019).”
Background
Metadata
- Authors
- Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., III, H. D., & Crawford, K.
- Affiliations
- DAIR Institute (US)
- University of Washington (US)
- Cornell University (US)
- Microsoft Research New York City (United States) (US)
- University of Maryland, College Park, MD
- University of Southern California (US)
- ORCID
- Venue
- Communications of the ACM
- Published
- 2021-11-19
- Volume
- 64
- Issue
- 12
- Pages
- 86-92
- ISSN
- 0001-0782 (ISSN-L)
- 1557-7317
- OpenAlex
- W2795038878
- Semantic Scholar
- 0df347f5e3118fac7c351917e3a497899b071d1e
- Cited by
- 2,469
- References
- 37
- Citations by year
- 2026: 788 · 2025: 402 · 2024: 298 · 2023: 379 · 2022: 221 · 2021: 217
- FWCI
- 252.30
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Communications of the ACM (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation
- Alfred P. Sloan Foundation
- Topics
- Physical Sciences › Computer Science › Information Systems › Research Data Management Practices
- Social Sciences › Decision Sciences › Information Systems and Management › Scientific Computing and Data Management
- Physical Sciences › Computer Science › Artificial Intelligence › Semantic Web and Ontologies
- Keywords
- dataset documentation, data provenance, data sharing, data reuse, metadata
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.