Data Statements for Natural Language Processing: Toward Mitigating System Bias and Enabling Better Science
Open accessThe DOI matches its Crossref record; 46 of the 46 references it cites were found in a registry or an index.
Abstract
In this paper, we propose data statements as a design solution and professional practice for natural language processing technologists, in both research and development. Through the adoption and widespread use of data statements, the field can begin to address critical scientific and ethical issues that result from the use of data from certain populations in the development of technology for other populations. We present a form that data statements can take and explore the implications of adopting them as part of regular practice. We argue that data statements will help alleviate issues related to exclusion and bias in language technology, lead to better precision in claims about how natural language processing research can generalize and thus better engineering results, protect companies from public embarrassment, and ultimately lead to language technology that meets its users in their own preferred linguistic style and furthermore does not misrepresent them to others.
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 cited source says what the claim attributes to it.
In the citing paper
“or lead to adverse effects from deploying suboptimal or harmful models (Bender and Friedman, 2018) on tasks for which they are ill-suited (Raji et al., 2022).”
From Lessons from the Trenches on Reproducible Evaluation of Language Models (2024) · cites this paper as Bender and Friedman, 2018
The cited source says what the claim attributes to it.
In the citing paper
“Numerous authors have observed that ML systems perform better for users whose demographic characteristics match those represented in the training data [9,36].”
From A Survey of Data Quality Requirements That Matter in ML Development Pipelines (2023) · cites this paper as [9]
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
Moher, D., Hopewell, S., Schulz, K. F., Montori, V., Gotzsche, P. C., Devereaux, P. J., Elbourne, D., Egger, M., & Altman, D. G. (2010) · BMJ · doi:10.1136/bmj.c869 · Cited by 8,883
Building a Large Annotated Corpus of English: The Penn Treebank
Marcus, M. (1993) · doi:10.21236/ada273556 · Cited by 7,405
Thumbs up?: sentiment classification using machine learning techniques
Pang, B., Lee, L., & Vaithyanathan, S. (2002) · Proceedings of the ACL-02 conference on Empirical methods in natural language processing - EMNLP '02 · doi:10.3115/1118693.1118704 · Cited by 7,024 · cited 1×
“Here, our focus is on linguistic data. the review (e.g., Pang et al., 2002) or the hashtag #sarcasm used to identify sarcastic language (e.g., Kreuz and Caucci, 2007).”
BackgroundThe Study of Second Language Acquisition
Brown, C., & Ellis, R. (1996) · The Modern Language Journal · doi:10.2307/329065 · Cited by 4,199
Publication Manual of the American Psychological Association
DeCleene, K. E., & Fogo, J. (2012) · Occupational Therapy In Health Care · doi:10.3109/07380577.2011.629024 · Cited by 3,167
The Social Stratification of English in New York City.
Crockett, H. J., & Labov, W. (1968) · American Sociological Review · doi:10.2307/2092899 · Cited by 3,084
Accurate unlexicalized parsing
Klein, D., & Manning, C. D. (2003) · Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - ACL '03 · doi:10.3115/1075096.1075150 · Cited by 3,008
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé, H., & Crawford, K. (2018) · arXiv · doi:10.48550/arxiv.1803.09010 · Cited by 2,474
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., III, H. D., & Crawford, K. (2021) · Communications of the ACM · doi:10.1145/3458723 · Cited by 2,469
Hateful Symbols or Hateful People? Predictive Features for Hate Speech Detection on Twitter
Waseem, Z., & Hovy, D. (2016) · Proceedings of the NAACL Student Research Workshop · doi:10.18653/v1/n16-2013 · Cited by 1,655 · cited 4×
“Waseem and Hovy (2016) provide the following complete list of terms used in their initial scrape: MKR, asian drive, feminazi, immigrant, nigger, sjw, WomenAgainstFeminism, blameonenotall, islam terrorism, notallmen, victimcard, victim card, arab terror, gamergate, jsil, racecard, race card. dataset…”
BackgroundMethodInfluential
Cited by
Items 101–200 of 865
Razi, A., Kim, S., Alsoubai, A., Stringhini, G., Solorio, T., De Choudhury, M., & Wisniewski, P. J. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3479609 · Cited by 95
Toward a Chatbot for Financial Sustainability
Hwang, S., & Kim, J. (2021) · Sustainability · doi:10.3390/su13063173 · Cited by 95
“Natural language processing technology and speech recognition technology are currently providing personal assistant services that communicate directly through personal mobile devices [4].”
BackgroundHow We Do Things With Words: Analyzing Text as Social and Cultural Data
Nguyen, D., Liakata, M., DeDeo, S., Eisenstein, J., Mimno, D., Tromble, R., Winters, J., Processing, S. N. L., & Processing, N. L. (2020) · Utrecht University Repository (Utrecht University) · Cited by 92
An Expert Annotated Dataset for the Detection of Online Misogyny
Guest, E., Vidgen, B., Mittos, A., Sastry, N. R., Tyson, G., & Margetts, H. Z. (2021) · Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/2021.eacl-main.114 · Cited by 181
HONEST: Measuring Hurtful Sentence Completion in Language Models
Nozza, D., Bianchi, F., & Hovy, D. (2021) · Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2021.naacl-main.191 · Cited by 95
“We follow Bender and Friedman (2018) on providing a Data Statement for our templates to provide a better picture of the possibilities and limitations of the data, and to allow future researchers to spot any biases we might have missed.”
MethodMoral Stories: Situated Reasoning about Norms, Intents, Actions, and their Consequences
Emelin, D., Le Bras, R., Hwang, J. D., Forbes, M., & Choi, Y. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.54 · Cited by 89
“The following data statement (Bender and Friedman, 2018) summarizes relevant aspects of the data collection process:”
BackgroundBurr, C., & Leslie, D. (2021) · AI and Ethics · doi:10.1007/s43681-022-00178-0 · Cited by 147
“The problem here is not that tools and method such as Datasheets [27], Data Nutrition Labels [36], Data Statements [7], Model Cards [51], and FactSheets [3] are of no use as provisional attempts at closing the gap between principles and practice, but rather that the limited perspective they take is liable to neglect t…”
MethodA large-scale audit of dataset licensing and attribution in AI
Longpre, S., Mahari, R., Chen, A., Obeng-Marnu, N., Sileo, D., Brannon, W., Muennighoff, N., Khazam, N., Kabbara, J., Perisetla, K., et al. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00878-8 · Cited by 86
“Previous work has stressed the importance of data documentation and attribution 22,67 of these fields (rarely the case), it is challenging to cleanly synthesize comprehensive and navigable documentation for the resulting bundle.”
One Country, 700+ Languages: NLP Challenges for Underrepresented Languages and Dialects in Indonesia
Aji, A. F., Winata, G. I., Koto, F., Cahyawijaya, S., Romadhony, A., Mahendra, R., Kurniawan, K., Moeljadi, D., Prasojo, R. E., Baldwin, T., et al. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-long.500 · Cited by 262
Liao, Q. V., Subramonyam, H., Wang, J., & Wortman Vaughan, J. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3580652 · Cited by 86
“To facilitate efective and consistent AI transparency, the AI research community has proposed various frameworks for transparent reporting of data [15, 45, 53], models [78], and services [12], often broadly referred to as “AI documentation.”
BackgroundThe Sanction of Authority: Promoting Public Trust in AI
Knowles, B., & Richards, J. T. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445890 · Cited by 86
A Checklist to Combat Cognitive Biases in Crowdsourcing
Draws, T., Rieger, A., Inel, O., Gadiraju, U., & Tintarev, N. (2021) · Proceedings of the AAAI Conference on Human Computation and Crowdsourcing · doi:10.1609/hcomp.v9i1.18939 · Cited by 86
“In the NLP field, Bender and Friedman (2018) proposed data statements , a characterization for data sets that provides relevant details regarding the population involved in creating a given data set, how the data set is used in experimental work, and how potential biases in the data set might affect…”
MethodIntroducing CAD: the Contextual Abuse Dataset
Vidgen, B., Nguyen, D., Margetts, H. Z., Rossini, P., & Tromble, R. (2021) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2021.naacl-main.182 · Cited by 167
“See Data Statement in the Appendix for more information.”
BackgroundMethodInfluentialA Survey of Data Quality Requirements That Matter in ML Development Pipelines
Priestley, M., O’donnell, F., & Simperl, E. (2023) · Journal of Data and Information Quality · doi:10.1145/3592616 · Cited by 86
“[3, 5, 8, 9, 25, 38, 70] Clarity of the data preprocessing sequence can be improved using documentation and publication of code.”
BackgroundMethodInfluentialIdentifying and Measuring Annotator Bias Based on Annotators’ Demographic Characteristics
Al Kuwatly, H., Wich, M., & Groh, G. (2020) · Proceedings of the Fourth Workshop on Online Abuse and Harms · doi:10.18653/v1/2020.alw-1.21 · Cited by 83
“Therefore, it is not surprising that a large body of work has identified and mitigated this bias (Bender and Friedman, 2018; Bountouridis et al., 2019; Dixon et al., 2018).”
BackgroundDocumenting Computer Vision Datasets: An Invitation to Reflexive Data Practices
Miceli, M., Yang, T., Naudts, L., Schuessler, M., Serbanescu, D., & Hanna, A. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445880 · Cited by 82
“Bender and Friedman [4] further suggest that such documentation should be required from “academics, industry, and government.”
BackgroundMethodInfluentialAlonso, M. Á., Vilares, D., Gómez‐Rodríguez, C., & Vilares, J. (2013) · Electronics · doi:10.3390/electronics · Cited by 83
A Human-Centered Systematic Literature Review of Cyberbullying Detection Algorithms
Kim, S., Razi, A., Stringhini, G., Wisniewski, P. J., & De Choudhury, M. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3476066 · Cited by 79
Markl, N. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533117 · Cited by 78
“Understanding datasets as infrastructure and pri-oritising good documentation is a first step to mitigating harms of biased systems as it allows us to anticipate them [15, 71, 99].”
BackgroundPublic procurement of artificial intelligence systems: new risks and future proofing
Hickok, M. (2024) · AI & SOCIETY · doi:10.1007/s00146-022-01572-2 · Cited by 81
“Such disclosures can be in the form of transparent procurement documentation, mandated human rights impact assessments, registries, as well as specification documents detailing the qualities of the datasets used and the design decisions embedded in the AI systems (Bender and Friedman 2018; Gebru et al. 2021; Hind et a…”
Improving fairness in machine learning systems: What do industry practitioners need?
Holstein, K., Vaughan, J. W., Daumé, H., Dudík, M., & Wallach, H. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.05239 · Cited by 50
Formalizing Trust in Artificial Intelligence: Prerequisites, Causes and Goals of Human Trust in AIPreprint
Jacovi, A., Marasović, A., Miller, T., & Goldberg, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.07487 · Cited by 81
“defining contracts is the work that proposes standardized documentations to communicate the performance characteristics of trained AI models. The examples of such documentations are: data statements [6], datasheets for datasets [18], model cards [54], reproducibility checklists [56], fairness checklists [48], and fact…”
BackgroundGender Bias in Word Embeddings: A Comprehensive Analysis of Frequency, Syntax, and Semantics
Caliskan, A., Ajay, P. P., Charlesworth, T., Wolfe, R., & Banaji, M. R. (2022) · Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3514094.3534162 · Cited by 79
“Furthermore, complementing computational methods by adding comprehensive data statements to language technologies as outlined by Bender and Friedman [5] will enhance transparency and raise awareness, which is the first step towards mitigating bias.”
MethodThe Gradient of Generative AI Release: Methods and Considerations
Solaiman, I. (2023) · 2023 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3593013.3593981 · Cited by 90
“Data statements for natural language processing [5] are another popular tool more tailored to language-based systems, seen used by bias measurement dataset CrowS-Pairs [62].”
MethodHealthsheet: Development of a Transparency Artifact for Health Datasets
Rostamzadeh, N., Mincu, D., Roy, S., Smart, A., Wilcox, L., Pushkarna, M., Schrouff, J., Amironesei, R., Moorosi, N., & Heller, K. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533239 · Cited by 72
“As a step towards more accountable ML practices[28], several guidelines and frameworks have been proposed around data transparency, collection and use in ML [9, 10, 22, 23, 49, 56].”
BackgroundOn Releasing Annotator-Level Labels and Information in Datasets
Prabhakaran, V., Mostafazadeh Davani, A., & Diaz, M. (2021) · Proceedings of The Joint 15th Linguistic Annotation Workshop (LAW) and 3rd Designing Meaning Representations (DMR) Workshop · doi:10.18653/v1/2021.law-1.14 · Cited by 75
The “Problem” of Human Label Variation: On Ground Truth in Data, Modeling and Evaluation
Plank, B. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.731 · Cited by 121
“More-over, if possible to release responsibly, besides making data statements of datasets available (Bender and Friedman, 2018), we encourage the community to include annotator-level background information (Prabhakaran et al., 2021) and document the annotation process (Geiger et al., 2020).”
BackgroundThompson, H. M., Sharma, B., Bhalla, S., Boley, R., McCluskey, C., Dligach, D., Churpek, M. M., Karnik, N. S., & Afshar, M. (2021) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocab148 · Cited by 78
“In machine learning, bias impacts every step of model development, testing, and implementation and can lead to algorithmic bias and feedback loops known as biased network effects.(7) Biases tend to unfairly disadvantage some groups or populations over others—often those already disproportionately marginalized.”
BackgroundDíaz, M., Kivlichan, I., Rosen, R., Baker, D., Amironesei, R., Prabhakaran, V., & Denton, R. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3534647 · Cited by 72
“CrowdWorkSheets complements and extends dataset development and documentation frameworks that have previously been developed in service of transparency, accountability, and reproducibility [6, 9, 17, 26, 27, 40, 42], but focuses specifically on unique considerations relating to crowdsourced dataset annotation.”
How We Do Things With Words: Analyzing Text as Social and Cultural Data
Nguyen, D., Liakata, M., DeDeo, S., Eisenstein, J., Mimno, D., Tromble, R., & Winters, J. (2020) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2020.00062 · Cited by 165
“These questions complement recent work providing guidance and suggestions about datasets (Bender and Friedman, 2018; Gebru et al., 2018) and models (Mitchell et al.”
Sharif, O., & Hoque, M. M. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.12.022 · Cited by 84
What’s in the Box? An Analysis of Undesirable Content in the Common Crawl Corpus
Luccioni, A., & Viviano, J. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers) · doi:10.18653/v1/2021.acl-short.24 · Cited by 69
“…high quality and safe neural language models will likely require the community to adopt more mindful data collection practices (Gehman et al., 2020; Bender and Friedman, 2018; Gebru et al., 2018; Jo and Gebru, 2020; Paullada et al., 2020; Bender et al., 2021), establish standardized filtering…”
BackgroundDo large language models have a legal duty to tell the truth?
Wachter, S., Mittelstadt, B., & Russell, C. (2024) · Royal Society Open Science · doi:10.1098/rsos.240197 · Cited by 66
“Instead, reliability can be improved through extensive curation and annotation [50], methodologically sound benchmarking metrics [51,52], long-term fine tuning with expert human feedback [53], auditing and adversarial testing ([5], p. 11, [53], p. 60), and perhaps even downsizing models [17, pp.…”
BackgroundPerturbation augmentation for fairer NLP
Qian, R., Ross, C., Fernandes, J., Smith, E. M., Kiela, D., & Williams, A. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.646 · Cited by 69
Toward Explainable Affective Computing: A Review
Cortiñas-Lorenzo, K., & Lacey, G. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3270027 · Cited by 111
Datasheets for Datasets help ML Engineers Notice and Understand Ethical Issues in Training Data
Boyd, K. L. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3479582 · Cited by 67
“Others take another approach: in their paper proposing Data Statements for Natural Language Processing, Bender & Friedman argue that their proposed context document may surface bias and other ethical problems without a question directly about ethics: “We propose here that foregrounding the characteristics of our datas…”
BackgroundInfluentialStance Prediction and Claim Verification: An Arabic Perspective
Khouja, J. (2020) · Fact Extraction and VERification Workshop (FEVER) · doi:10.18653/v1/2020.fever-1.2 · Cited by 96
SkillSpan: Hard and Soft Skill Extraction from English Job Postings
Zhang, M., Jensen, K. N., Sonniks, S. D., & Plank, B. (2022) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2022.naacl-main.366 · Cited by 140
Evaluation Gaps in Machine Learning Practice
Hutchinson, B., Rostamzadeh, N., Greer, C., Heller, K., & Prabhakaran, V. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533233 · Cited by 62
Stalnaker, T., Wintersgill, N., Chaparro, O., Penta, M. D., Germán, D. M., & Poshyvanyk, D. (2024) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1145/3597503.3623347 · Cited by 129
“There are different proposals to track datasets and AI model information [43, 66, 72, 96].”
BackgroundProblems with Cosine as a Measure of Embedding Similarity for High Frequency Words
Zhou, K., Ethayarajh, K., Card, D., & Jurafsky, D. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-short.45 · Cited by 99
Bias in Automatic Speech Recognition: The Case of African American Language
Martin, J. L., & Wright, K. E. (2023) · Applied Linguistics · doi:10.1093/applin/amac066 · Cited by 65
Assessing the Impact of Contextual Information in Hate Speech Detection
Pérez, J. M., Luque, F. M., Zayat, D., Kondratzky, M., Moro, A., Serrati, P. S., Zajac, J., Miguel, P., Debandi, N., Gravano, A., et al. (2023) · IEEE Access · doi:10.1109/access.2023.3258973 · Cited by 115
“Following the lines of Data Statements [61], we provide in this subsection a profile of the annotators.”
MethodUsing ChatGPT in Education: Human Reflection on ChatGPT’s Self-Reflection
Loos, E., Gröpler, J., & Goudeau, M.-L. S. (2023) · Societies · doi:10.3390/soc13080196 · Cited by 64
“They also result from prompts that are abstract or contain incorrect information [10,15–20].”
Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition
Huang, X., Xing, L., Dernoncourt, F., & Paul, M. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.10361 · Cited by 21
“Our code, anonymized data and data statement (Bender and Friedman, 2018) will be publicly available at https://github.”
MethodA pre-trained BERT for Korean medical natural language processing
Kim, Y., Kim, J.-H., Lee, J. M., Jang, M. J., Yum, Y., Kim, S., Shin, U., Kim, Y.-M., Joo, H. J., & Song, S. (2022) · Scientific Reports · doi:10.1038/s41598-022-17806-8 · Cited by 77
Data and its (dis)contents: A survey of dataset development and use in machine learning research
Paullada, A., Raji, I. D., Bender, E. M., Denton, E., & Hanna, A. (2021) · Patterns · doi:10.1016/j.patter.2021.100336 · Cited by 548
Self-citationYou reap what you sow: On the Challenges of Bias Evaluation Under Multilingual Settings
Talat, Z., Névéol, A., Biderman, S., Clinciu, M., Dey, M., Longpre, S., Luccioni, S., Masoud, M., Mitchell, M., Radev, D., et al. (2022) · Proceedings of BigScience Episode #5 -- Workshop on Challenges & Perspectives in Creating Large Language Models · doi:10.18653/v1/2022.bigscience-1.3 · Cited by 59
“Such issues are in part ex-186 acerbated by the general poverty of documentation of datasets (Gebru et al., 2018; Bender and Friedman, 2018) and machine learning models (Mitchell et al., 2019).”
BackgroundFairness in Large Language Models: A Taxonomic Survey
Chu, Z., Wang, Z., & Zhang, W. (2024) · ACM SIGKDD Explorations Newsletter · doi:10.1145/3682112.3682117 · Cited by 59
Moral concerns are differentially observable in language
Kennedy, B., Atari, M., Davani, A. M., Hoover, J., Omrani, A., Graham, J., & Dehghani, M. (2021) · Cognition · doi:10.1016/j.cognition.2021.104696 · Cited by 91
Heger, A. K., Marquis, L. B., Vorvoreanu, M., Wallach, H., & Wortman Vaughan, J. (2022) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3555760 · Cited by 58
“Proposed data documentation frameworks include datasheets for datasets [19], dataset nutrition labels [25], and data statements for natural language processing datasets [7].”
BackgroundInfluentialTo protect science, we must use LLMs as zero-shot translators
Mittelstadt, B., Wachter, S., & Russell, C. (2023) · Nature Human Behaviour · doi:10.1038/s41562-023-01744-0 · Cited by 58
Fairness in Deep Learning: A Survey on Vision and Language Research
Parraga, O., Móre, M. D., Mattjie, C., Gavenski, N. S., Kupssinskü, L. S., Medronha, A., de Moura, L. V., Simões, G. S., & Barros, R. C. (2023) · ACM Computing Surveys · doi:10.1145/3637549 · Cited by 123
“There are several alternative approaches for documenting datasets; two examples include the Dataset Nutrition Label [81] and Data Statements for NLP [9].”
Talk2Nav: Long-Range Vision-and-Language Navigation with Dual Attention and Spatial Memory
Vasudevan, A. B., Dai, D., & Gool, L. V. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-020-01374-3 · Cited by 101
Making It Possible for the Auditing of AI: A Systematic Review of AI Audits and AI Auditability
Li, Y., & Goel, S. (2024) · Information Systems Frontiers · doi:10.1007/s10796-024-10508-8 · Cited by 105
“The standardized documentation for data (i.e., Data Statements (Bender & Friedman, 2018), Datasheets (Gebru et al., 2021), and Data Nutrition Labels (Holland et al., 2020; Chmielinski et al., 2022) and AI models (i.e., Model Cards (Mitchell et al., 2019), Interactive Model Cards (Crisan et al.,…”
InfluentialReturning the N to NLP: Towards Contextually Personalized Classification Models
Flek, L. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.700 · Cited by 56
“As pointed out in Bender and Friedman (2018), we rarely get to know anything about the people other than if they were “expert” 1 .”
BackgroundTeach Me to Explain: A Review of Datasets for Explainable NLP.Preprint
Wiegreffe, S., & Marasović, A. (2021) · arXiv (Cornell University) · Cited by 64
Session-based cyberbullying detection in social media: A survey
Yi, P., & Zubiaga, A. (2023) · Online Social Networks and Media · doi:10.1016/j.osnem.2023.100250 · Cited by 55
Perez-Almendros, C., Espinosa-Anke, L., & Schockaert, S. (2020) · International Conference on Computational Linguistics · doi:10.18653/v1/2020.coling-main.518 · Cited by 85
Razi, A., Alsoubai, A., Kim, S., Naher, N., Ali, S., Stringhini, G., De Choudhury, M., & Wisniewski, P. J. (2022) · CHI Conference on Human Factors in Computing Systems Extended Abstracts · doi:10.1145/3491101.3503569 · Cited by 53
“To make online risk detection systems timely, scalable, and most importantly, accurate, it is crucial that the detection models are built upon ecologically valid datasets that depict the target users (i.e., youth) [10].”
BackgroundTeach Me to Explain: A Review of Datasets for Explainable Natural Language ProcessingPreprint
Wiegreffe, S., & Marasović, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.12060 · Cited by 149
“It is important to constantly evaluate data collection practices critically and standardize them [13, 39, 95].”
BackgroundLu, J. H., Callahan, A., Patel, B. S., Morse, K. E., Dash, D., Pfeffer, M. A., & Shah, N. H. (2022) · JAMA Network Open · doi:10.1001/jamanetworkopen.2022.27779 · Cited by 52
El–Sappagh, S., Alonso, J. M., Abuhmed, T., Ali, F., & Bugarín, A. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10415-5 · Cited by 69
Changing the World by Changing the Data
Rogers, A. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.170 · Cited by 51
“has to be thorough documentation of the data (Bender and Friedman, 2018; Gebru et al., 2020): this lets us compare the represented population and the population of the target users, and think through the possible harms.”
BackgroundIntrinsic Evaluation of Summarization Datasets
Bommasani, R., & Cardie, C. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.649 · Cited by 51
“Data understanding can also inform scientific and ethical decision-making (Bender and Friedman, 2018; Gebru et al., 2018; Mitchell et al., 2019) with recent work studying how social biases encoded in”
BackgroundBad Seeds: Evaluating Lexical Methods for Bias Measurement
Antoniak, M., & Mimno, D. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.148 · Cited by 50
A participatory data-centric approach to AI Ethics by Design
Gerdes, A. (2021) · Applied Artificial Intelligence · doi:10.1080/08839514.2021.2009222 · Cited by 77
A hunt for the Snark: Annotator Diversity in Data Practices
Kapania, S., Taylor, A. S., & Wang, D. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3580645 · Cited by 50
“, Datasheets for Datasets [30], Data Cards [77], Data Statements [5]) and archival artefacts of the decisions made [48] during the annotation process were proposed as a mechanism to foster deliberative accountability [74], to make explicit the tacit knowledge in data work [63, 73] and to foster fair reparation in ML b…”
Khurram, M., Zhang, C., Muhammad, S., Kishnani, H., An, K. R., Abeywardena, K. G., Chadha, U., & Behdinan, K. (2025) · Processes · doi:10.3390/pr13051312 · Cited by 90
The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics
Gehrmann, S., Adewumi, T., Aggarwal, K., Ammanamanchi, P. S., Anuoluwapo, A., Bosselut, A., Chandu, K. R., Clinciu, M., Das, D., Dhole, K. D., et al. (2021) · arXiv · doi:10.48550/arxiv.2102.01672 · Cited by 49
“cards (Bender and Friedman, 2018; Gebru et al., 2018) we developed and for which we release a template and guide.”
MethodMitigating Gender Bias in Captioning Systems
Tang, R., Du, M., Li, Y., Liu, Z., Zou, N., & Hu, X. (2021) · Proceedings of the Web Conference 2021 · doi:10.1145/3442381.3449950 · Cited by 50
Symphony: Composing Interactive Interfaces for Machine Learning
Bäuerle, A., Cabrera, Á. A., Hohman, F., Maher, M., Koski, D., Suau, X., Barik, T., & Moritz, D. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3502102 · Cited by 69
“Similar work has focused on specific types of data, for example, Data Statements [7] are tailored to natural language processing datasets.”
BackgroundAnalyzing Dataset Annotation Quality Management in the Wild
Klie, J.-C., de Castilho, R. E., & Gurevych, I. (2024) · Computational Linguistics · doi:10.1162/coli_a_00516 · Cited by 136
“While not ideal, this is a similar setup as used in previous works surveying NLP publications (Sabou et al. 2014; Amidei, Piwek, and Willis 2019; Dror et al. 2018; Shmueli et al. 2021).”
BackgroundInfluentialDehghani, M., Tay, Y., Gritsenko, A. A., Zhao, Z., Houlsby, N., Diaz, F., Metzler, D., & Vinyals, O. (2021) · arXiv · doi:10.48550/arxiv.2107.07002 · Cited by 48
Cao, Y. T., & Daumé, H. (2021) · Computational Linguistics · doi:10.1162/coli_a_00413 · Cited by 47
Röttger, P., & Pierrehumbert, J. B. (2021) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2021.findings-emnlp.206 · Cited by 102
Evaluation Examples are not Equally Informative: How should that change NLP Leaderboards?
Rodriguez, P., Barrow, J., Hoyle, A. M., Lalor, J. P., Jia, R., & Boyd-Graber, J. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.346 · Cited by 47
“Leaderboards are also well positioned to provide incentive structures for participants to prioritize fairness (Bender and Friedman, 2018) and efficiency (Strubell et al.”
BackgroundThe unbearable (technical) unreliability of automated facial emotion recognition
Cabitza, F., Campagner, A., & Mattioli, M. (2022) · Big Data & Society · doi:10.1177/20539517221129549 · Cited by 70
“More in general, our results resonate with recent initiatives (Bender and Friedman, 2018; Gebru et al., 2021; Holland et al., 2018) aimed at raising awareness about the data production process (Gitelman, 2013), including the need to document in which (technical, social, economical and political)…”
ResultThe Foundation Model Transparency Index
Bommasani, R., Klyman, K., Longpre, S., Kapoor, S., Maslej, N., Xiong, B., Zhang, D., & Liang, P. (2023) · arXiv · doi:10.48550/arxiv.2310.12941 · Cited by 46
Data Governance in the Age of Large-Scale Data-Driven Language Technology
Jernite, Y., Nguyen, H., Biderman, S., Rogers, A., Masoud, M., Danchev, V., Tan, S., Luccioni, A. S., Subramani, N., Johnson, I., et al. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3534637 · Cited by 75
Multilingual HateCheck: Functional Tests for Multilingual Hate Speech Detection Models
Röttger, P., Seelawi, H., Nozza, D., Talat, Z., & Vidgen, B. (2022) · Workshop on Online Abuse and Harms (WOAH) · doi:10.18653/v1/2022.woah-1.15 · Cited by 118
Mittal, S., Thakral, K., Singh, R., Vatsa, M., Glaser, T., Ferrer, C. C., & Hassner, T. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00874-y · Cited by 79
“In the field of NLP, researchers have proposed data statements to understand bias and intent 15 .”
Kirk, H., Vidgen, B., Rottger, P., Thrush, T., & Hale, S. (2022) · Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2022.naacl-main.97 · Cited by 43
Embeddings in Natural Language Processing: Theory and Advances in Vector Representations of Meaning
Pilehvar, M. T., & Camacho-Collados, J. (2020) · Synthesis lectures on human language technologies · doi:10.2200/s01057ed1v01y202009hlt047 · Cited by 88
Zeinert, P., Inie, N., & Derczynski, L. (2021) · doi:10.18653/v1/2021.acl-long.247 · Cited by 88
“3.1 Dataset: Online misogyny in social media Bender and Friedman (2018) present a set of data statements for NLP which help “alleviate issues re-lated to exclusion and bias in language technology, lead[ing] to better precision in claims about how natural language processing research can generalize…”
Influential‘Just What do You Think You’re Doing, Dave?’ A Checklist for Responsible Data Use in NLP
Rogers, A., Baldwin, T., & Leins, K. (2021) · Findings of the Association for Computational Linguistics: EMNLP 2021 · doi:10.18653/v1/2021.findings-emnlp.414 · Cited by 53
Exploring factors influencing educators’ adoption of ChatGPT: a mixed method approach
Hidayat-ur-Rehman, I., & Ibrahim, Y. (2023) · Interactive Technology and Smart Education · doi:10.1108/itse-07-2023-0127 · Cited by 65
Characterizing and Detecting Mismatch in Machine-Learning-Enabled Systems
Lewis, G. A., Bellomo, S., & Özkaya, İ. (2021) · IEEE/ACM International Conference on AI Engineering – Software Engineering for AI (CAIN) · doi:10.1109/wain52551.2021.00028 · Cited by 58
Towards a multi-stakeholder value-based assessment framework for algorithmic systems
Yurrita, M., Murray-Rust, D., Balayn, A., & Bozzon, A. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533118 · Cited by 39
“For NLP techniques, “Data Statements” are regarded as a dataset characterization approach that helps developers anticipate biases in language technology and understand how these can be better deployed [20].”
Just What do You Think You're Doing, Dave?' A Checklist for Responsible Data Use in NLP
Rogers, A., Baldwin, T., & Leins, K. (2021) · arXiv · doi:10.48550/arxiv.2109.06598 · Cited by 42
“Creation of ‘perfect’ and representative samples is not necessarily possible, but documenting the lacunae and omissions is (Bender and Friedman, 2018).”
BackgroundMethodTan, M., & Subramonyam, H. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642592 · Cited by 40
CoMPosT: Characterizing and Evaluating Caricature in LLM Simulations
Cheng, M., Piccardi, T., & Yang, D. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.669 · Cited by 196
Berman, G., Goyal, N., & Madaio, M. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642398 · Cited by 40
“In this context, we conceive of a dataset transparency protocol as an RAI tool designed to enable data scientists to document relevant aspects of a dataset in a standardized manner [e.g., 22, 77, 162].”
MethodHopkins, A., & Booth, S. (2021) · Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3461702.3462527 · Cited by 40
“Nonetheless, we see this as an area which lacks sufficient investment both in research and in practice, and which could be improved through the introduction of comprehensive standards [4, 23, 34].”
BackgroundConvAbuse: Data, Analysis, and Benchmarks for Nuanced Detection in Conversational AI
Cercas Curry, A., Abercrombie, G., & Rieser, V. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.587 · Cited by 39
Softcite dataset: A dataset of software mentions in biomedical and economic research publications
Du, C., Cohoon, J., Lopez, P., & Howison, J. (2021) · Journal of the Association for Information Science and Technology · doi:10.1002/asi.24454 · Cited by 71
“” Some recommend documenting the demo-graphics of annotators (Bender & Friedman, 2018) as possibly affecting the bias and accountability of the dataset; however, should we discuss the immigration status of annotators, even after being made anonymous, as we did above?”
BackgroundMethodPredicting Performance for Natural Language Processing Tasks
Xia, M., Anastasopoulos, A., Xu, R., Yang, Y., & Neubig, G. (2020) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2020.acl-main.764 · Cited by 83
FEEL-IT: Emotion and Sentiment Classification for the Italian Language
Bianchi, F., Nozza, D., & Hovy, D. (2021) · Archivio istituzionale della Ricerca - Bocconi · Cited by 37
Natural language processing for clinical notes in dentistry: A systematic review
Pethani, F., & Dunn, A. G. (2023) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2023.104282 · Cited by 54
McMillan-Major, A., Osei, S., Rodriguez, J. D., Ammanamanchi, P. S., Gehrmann, S., & Jernite, Y. (2021) · Proceedings of the 1st Workshop on Natural Language Generation, Evaluation, and Metrics (GEM 2021) · doi:10.18653/v1/2021.gem-1.11 · Cited by 37
“The first such example is by Bender and Friedman (2018) who propose a version of Data Statements for documenting aspects of the data from a linguistic perspective.”
BackgroundMethod
Metadata
- Authors
- Bender, E. M., & Friedman, B.
- Affiliations
- University of Washington (US)
- Corresponding author
- Bender, Emily M.
- Friedman, Batya
- Venue
- Transactions of the Association for Computational Linguistics
- Publisher
- MIT Press
- Published
- 2018-12
- Volume
- 6
- Pages
- 587-604
- ISSN
- 2307-387X (ISSN-L)
- OpenAlex
- W2911227954
- Semantic Scholar
- 97bfa89addc6e5d76361e4c1e296949cad887b86
- Cited by
- 854
- References
- 46
- Citations by year
- 2026: 108 · 2025: 96 · 2024: 102 · 2023: 154 · 2022: 127 · 2021: 177 · 2020: 68 · 2019: 20 · 2018: 1
- FWCI
- 132.62
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- cc-by
- Locations
- Transactions of the Association for Computational Linguistics (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Funders
- National Science Foundation (1302709)
- University of Washington
- York University
- Macquarie University
- University of California, San Diego
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Keywords
- natural language processing, algorithmic bias, underrepresented groups, ethical practice, linguistic diversity
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.