On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Open accessThe DOI matches its Crossref record; 156 of the 159 references it cites were found in a registry or an index, and 3 are not checked yet.
Abstract
The past 3 years of work in NLP have been characterized by the development and deployment of ever larger language models, especially for English. BERT, its variants, GPT-2/3, and others, most recently Switch-C, have pushed the boundaries of the possible both through architectural innovations and through sheer size. Using these pretrained models and the methodology of fine-tuning them for specific tasks, researchers have extended the state of the art on a wide array of tasks as measured by leaderboards on specific benchmarks for English. In this paper, we take a step back and ask: How big is too big? What are the possible risks associated with this technology and what paths are available for mitigating those risks? We provide recommendations including weighing the environmental and financial costs first, investing resources into curating and carefully documenting datasets rather than ingesting everything on the web, carrying out pre-development exercises evaluating how the planned approach fits into research and development goals and supports stakeholder values, and encouraging research directions beyond ever larger language models.
Checked 2026-11-22 · 3 references not checked yet
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The text we found for this paper looks like another document, so the self-check wasn’t run.
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 5×
“[39] in particular noted that training on a large dataset and fine-tuning for specific tasks leads to strictly increasing results on the GLUE tasks [138] for English as the hyperparameters of the model were increased.”
BackgroundMethodResultInfluentialGlove: Global Vectors for Word Representation
Pennington, J., Socher, R., & Manning, C. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1162 · Cited by 34,071
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 4×
“documents linked to from Reddit [25], plus Wikipedia and a collection”
BackgroundMethodInfluentialRoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515
Distilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204 · cited 1×
“[26, 58], quantization [118, 153], factorized embedding parameterization and cross-layer parameter sharing [70], and progressive module replacing [146].”
Deep Contextualized Word Representations
Peters, M., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1202 · Cited by 11,983
Proceedings of the 24th international conference on Machine learning
(2007) · doi:10.1145/1273496 · Cited by 11,630
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929
The Mathematical Theory of Communication
Shannon, C. E., Weaver, W., & Wiener, N. (1950) · Physics Today · doi:10.1063/1.3067010 · Cited by 9,801
Research works printed but not linked: 48
- [4] American Psychological Association. 2019. Discrimination: What it is, and how to cope. https://www.apa.org/topics/discrimination (2019). Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- [15] Ruha Benjamin. 2019. Race After Technology: Abolitionist Tools for the New Jim Code. Polity Press, Cambridge, UK. Crossref Google Scholar
- [16] Elettra Bietti and Roxana Vatanparast. 2020. Data Waste. Harvard International Law Journal 61 (2020). Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.1016/0010-0277(86) · doi:10.1016/0010-0277(86
- [38] Terrance de Vries, Ishan Misra, Changhan Wang, and Laurens van der Maaten. 2019. Does object recognition work for everyone?. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 52--59. Google Scholar
- 10.18653/v1 · DOI
- [42] Ethan Fast, Tina Vachovsky, and Michael Bernstein. 2016. Shirtless and Dangerous: Quantifying Linguistic Signals of Gender Bias in an Online Fiction Writing Community. In Proceedings of the International AAAI Conference on Web and Social Media, Vol. 10. Google Scholar
- 10.18653/v1 · DOI
- [45] Darja Fišer, Ruihong Huang, Vinodkumar Prabhakaran, Rob Voigt, Zeerak Waseem, and Jacqueline Wernimont (Eds.). 2018. Proceedings of the 2nd Workshop on Abusive Language Online (ALW2). Association for Computational Linguistics, Brussels, Belgium. https://www.aclweb.org/anthology/W18-5100 Google Scholar
- [50] Batya Friedman, Peter H. Kahn, Jr., and Alan Borning. 2006. Value sensitive design and information systems. In Human-Computer Interaction in Management Information Systems: Foundations, P Zhang and D Galletta (Eds.). M. E. Sharpe, Armonk NY, 348--372. Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- [67] Brendan Kennedy, Drew Kogon, Kris Coombs, Joseph Hoover, Christina Park, Gwenyth Portillo-Wightman, Aida Mostafazadeh Davani, Mohammad Atari, and Morteza Dehghani. 2018. A typology and coding manual for the study of hate-based rhetoric. PsyArXiv. July 18 (2018). Google Scholar
- [77] Marianna Martindale and Marine Carpuat. 2018. Fluency Over Adequacy: A Pilot Study in Measuring User Trust in Imperfect MT. In Proceedings of the 13th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Track). Association for Machine Translation in the Americas, Boston, MA, 13--25. https://www.aclweb.org/anthology/W18-1803 Google Scholar
- [81] Douglas M McLeod. 2007. News coverage and social protest: How the media's protect paradigm exacerbates social conflict. Journal of Dispute Resolution (2007), 185. Google Scholar
- [87] Robert C. Moore and William Lewis. 2010. Intelligent Selection of Language Model Training Data. In Proceedings of the ACL 2010 Conference Short Papers. Association for Computational Linguistics, Uppsala, Sweden, 220--224. https://www.aclweb.org/anthology/P10-2041 Google Scholar
- 10.18653/v1 · DOI
- [92] Maggie Nelson. 2015. The Argonauts. Graywolf Press, Minneapolis. Google Scholar
- 10.18653/v1 · DOI
- [96] David Ortiz, Daniel Myers, Eugene Walls, and Maria-Elena Diaz. 2005. Where do we stand with newspaper data? Mobilization: An International Quarterly 10, 3 (2005), 397--419. Google Scholar
- 10.3115/v1 · doi:10.3115/v1
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- [106] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI Blog 1, 8 (2019), 9. Google Scholar
- 10.18653/v1 · DOI
- [109] Sarah T. Roberts, Joel Tetreault, Vinodkumar Prabhakaran, and Zeerak Waseem (Eds.). 2019. Proceedings of the Third Workshop on Abusive Language Online. Association for Computational Linguistics, Florence, Italy. https://www.aclweb.org/anthology/W19-3500 Google Scholar
- [112] Corby Rosset. 2020. Turing-NLG: A 17-billion-parameter language model by Microsoft. Microsoft Blog (2020). Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- [121] Joonbo Shin, Yoonhyung Lee, and Kyomin Jung. 2019. Effective Sentence Scoring Method Using BERT for Speech Recognition. In Asian Conference on Machine Learning. 1081--1093. Google Scholar
- [124] Karen Spärck Jones. 2004. Language modelling's generative model: Is it rational? Technical Report. Computer Laboratory, University of Cambridge. Google Scholar
- [127] Katrina Srigley and Lorraine Sutherland. 2019. Decolonizing, Indigenizing, and Learning Biskaaybiiyang in the Field: Our Oral History Journey1. The Oral History Review (2019). Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- [150] Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Russ R Salakhutdinov, and Quoc V Le. 2019. Xlnet: Generalized autoregressive pretraining for language understanding. In Advances in neural information processing systems. 5753--5763. Google Scholar
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
- 10.18653/v1 · DOI
Cited by
Items 701–800 of 6,931
Ethics of generative AI and manipulation: a design-oriented research agenda
Klenk, M. (2024) · Ethics and Information Technology · doi:10.1007/s10676-024-09745-x · Cited by 75
“Notably, the indifference criterion can capture emergent, unwitting manipulation resulting from the sense that generative AI systems act as “stochastic parrots” (Bender et al., 2021).”
Metzger, L., Miller, L., Baumann, M., & Kraus, J. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642122 · Cited by 73
Large language models present new questions for decision support
Handler, A., Larsen, K. R. T., & Hackathorn, R. D. (2024) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2024.102811 · Cited by 108
Sinha, R., Solola, I., Nguyen, H., Swanson, H., & Lawrence, L. (2024) · Symposium on Learning, Design and Technology (LDT) · doi:10.1145/3663433.3663456 · Cited by 72
“Also, concerns about transparency, authorship, and accountability are ongoing areas of research when incorporating GenAI models into study designs [2, 16].”
MethodStrong and weak alignment of large language models with human values
Khamassi, M., Nahon, M., & Chatila, R. (2024) · Scientific Reports · doi:10.1038/s41598-024-70031-3 · Cited by 101
Explaining AI through mechanistic interpretability
Kästner, L., & Crook, B. (2024) · European Journal for Philosophy of Science · doi:10.1007/s13194-024-00614-4 · Cited by 79
“Other critics may worry that (#6) MI will not yield accurate predictions of AI systems’ behaviours in social contexts – though this is what really matters when we think about deploying AI in modern society (Bender et al., 2021).”
Knowledge Management in a World of Generative AI: Impact and Implications
Storey, V. C. (2025) · ACM Transactions on Management Information Systems · doi:10.1145/3719209 · Cited by 82
“At the same time, decades of research on knowledge management should be useful for understanding GenAI.”
Arar, K. H., Özen, H., Polat, G., & Turan, S. (2025) · Smart Learning Environments · doi:10.1186/s40561-025-00403-3 · Cited by 82
“It illuminates such important concepts as black box problem and algorithmic bias (Bender et al., 2021) and broadens normative ethical discussions of the morality of AI (Solum, 1992) and questions of deontological ethical dilemmas (Ditto & Liu, 2012).”
Agentic LLMs in the supply chain: towards autonomous multi-agent consensus-seeking
Jannelli, V., Schoepf, S., Bickel, M., Netland, T. H., & Brintrup, A. M. (2025) · International Journal of Production Research · doi:10.1080/00207543.2025.2604311 · Cited by 124
The AI ethicist’s dilemma: fighting Big Tech by supporting Big Tech
Sætra, H. S., Coeckelbergh, M., & Danaher, J. (2021) · AI and Ethics · doi:10.1007/s43681-021-00123-7 · Cited by 58
“These problems could be related to, amongst other things, polarization [1], fake news [2], filter bubbles [3], echo chambers [4], changes in human cognition [5], discrimination and bias [6], and surveillance and challenges related to a lack of privacy [1].”
BackgroundMethodInfluentialTowards 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
Measuring Representational Harms in Image Captioning
Wang, A., Barocas, S., Laird, K., & Wallach, H. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533099 · Cited by 40
de Fine Licht, K. (2024) · Philosophy & Technology · doi:10.1007/s13347-024-00799-9 · Cited by 40
Knowledge graph-extended retrieval augmented generation for question answering
Linders, J., & Tomczak, J. M. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06885-5 · Cited by 40
“Given a sequence x 1: t , the model produces a probability distribution p ( x t +1 | x 1: t ) = softmax( z/T ) over its vocabulary, where z are the raw logits and T is a temperature parameter that controls randomness.”
Ma, Z., Mei, Y., & Su, Z. (2023) · arXiv · doi:10.48550/arxiv.2307.15810 · Cited by 40
The AI Doctor Is In: A Survey of Task-Oriented Dialogue Systems for Healthcare Applications
Valizadeh, M., & Parde, N. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-long.458 · Cited by 101
“Any system trained on large, weakly curated datasets may also learn unpleasant behaviors and amplify biases in the training data, in turn producing harmful consequences (Dinan et al., 2021; Bender et al., 2021).”
BackgroundLi, B., Bonk, C. J., & Kou, X. (2023) · International Journal of Computer-Assisted Language Learning and Teaching · doi:10.4018/ijcallt.326135 · Cited by 58
Algorithmic bias in artificial intelligence is a problem—And the root issue is power
Walker, R., Dillard‐Wright, J., & Iradukunda, F. (2023) · Nursing Outlook · doi:10.1016/j.outlook.2023.102023 · Cited by 45
Balayn, A., Yurrita, M., Yang, J., & Gadiraju, U. (2023) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3600211.3604674 · Cited by 64
Safeguarding human values: rethinking US law for generative AI’s societal impacts
Cheong, I., Caliskan, A., & Kohno, T. (2024) · AI and Ethics · doi:10.1007/s43681-024-00451-4 · Cited by 86
The Impact and Opportunities of Generative AI in Fact-Checking
Wolfe, R., & Mitra, T. (2024) · The 2024 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3630106.3658987 · Cited by 40
“Bender et al. [2] contend that training models on poorly specified textual data poses numerous ethical risks for downstream use.”
BackgroundGenerative AI Meets Open-Ended Survey Responses: Research Participant Use of AI and Homogenization
Zhang, S., Xu, J., & Alvero, A. (2025) · Sociological Methods & Research · doi:10.1177/00491241251327130 · Cited by 102
Survey of Cultural Awareness in Language Models: Text and Beyond
Pawar, S., Park, J., Jin, J., Arora, A., Myung, J., Yadav, S., Haznitrama, F. G., Song, I., Oh, A. H., & Augenstein, I. (2025) · Computational Linguistics · doi:10.1162/coli.a.14 · Cited by 222
The ethics of AI or techno-solutionism? UNESCO’s policy guidance on AI in education
Mochizuki, Y., Bruillard, É., & Bryan, A. (2025) · British Journal of Sociology of Education · doi:10.1080/01425692.2025.2502808 · Cited by 81
Hallucination to truth: a review of fact-checking and factuality evaluation in large language models
Rahman, S. S., Islam, M. A., Alam, M. M., Zeba, M., Rahman, M. A., Chowa, S. S., Raiaan, M. A. K., & Azam, S. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-025-11454-w · Cited by 39
Exploring AI-Generated Text in Student Writing: How Does AI Help?
Woo, D. J., Susanto, H., Yeung, C. H., Guo, K., & Fung, A. K. Y. (2023) · arXiv · doi:10.48550/arxiv.2304.02478 · Cited by 39
Language Generation Models Can Cause Harm: So What Can We Do About It? An Actionable Survey
Kumar, S., Balachandran, V., Njoo, L., Anastasopoulos, A., & Tsvetkov, Y. (2023) · Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/2023.eacl-main.241 · Cited by 38
Making and Using AI in the Library: Creating a BERT Model at the National Library of Sweden
Haffenden, C., Fano, E., Malmsten, M., & Börjeson, L. (2023) · College & Research Libraries · doi:10.5860/crl.84.1.30 · Cited by 62
GPThreats-3: Is Automatic Malware Generation a Threat?
Botacin, M. F. (2023) · IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw59333.2023.00027 · Cited by 66
“Microsoft’s model has 17 billion parameters, GPT-3 has 175 billion parameters, and GPT-4 is expected to reach 100 trillion parameters [34] (for full model information, see [2]).”
BackgroundThe Neuroscience of Human and Artificial Intelligence Presence
Harris, L. T. (2023) · Annual Review of Psychology · doi:10.1146/annurev-psych-013123-123421 · Cited by 49
Geotechnical Parrot Tales (GPT): Harnessing Large Language Models in Geotechnical Engineering
Kumar, K. (2023) · Journal of Geotechnical and Geoenvironmental Engineering · doi:10.1061/jggefk.gteng-11828 · Cited by 52
mGPT: Few-Shot Learners Go Multilingual
Shliazhko, O., Fenogenova, A., Tikhonova, M., Kozlova, A., Mikhailov, V., & Shavrina, T. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00633 · Cited by 277
“Last but not least, finetuning requires additional computational resources and, therefore, aggravates the problem of a large carbon footprint (Bender et al., 2021).”
BackgroundBhattacharjee, A., Zeng, Y., Xu, S. Y., Kulzhabayeva, D., Ma, M., Kornfield, R., Ahmed, S. I., Mariakakis, A. T., Czerwinski, M. P., Kuzminykh, A., et al. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642081 · Cited by 73
“While the promises of LLMs in academic settings are manifold, these benefits come with ethical concerns that cannot be ignored [7, 50, 52, 54, 74, 105].”
BackgroundThe implications of generative artificial intelligence for mathematics education
Walkington, C. (2025) · School Science and Mathematics · doi:10.1111/ssm.18356 · Cited by 100
The use of LLMs to annotate data in management research: Foundational guidelines and warnings
Carlson, N. A., & Burbano, V. (2025) · Strategic Management Journal · doi:10.1002/smj.70023 · Cited by 87
Névéol, A., Dupont, Y., Bezançon, J., & Fort, K. (2022) · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2022.acl-long.583 · Cited by 38
“The increasing use of large language models has raised many ethical concerns, including the risk of bias and bias amplification (Bender et al., 2021).”
BackgroundWang, L., & Ren, B. (2024) · Education Sciences · doi:10.3390/educsci14121329 · Cited by 38
GDPR and Large Language Models: Technical and Legal Obstacles
Feretzakis, G., Vagena, E., Kalodanis, K., Peristera, P., Kalles, D., & Anastasiou, A. (2025) · Future Internet · doi:10.3390/fi17040151 · Cited by 38
Moon, K., Green, A. E., & Kushlev, K. (2025) · Computers in Human Behavior: Artificial Humans · doi:10.1016/j.chbah.2025.100207 · Cited by 38
Synthetic Replacements for Human Survey Data? The Perils of Large Language Models
Bisbee, J., Clinton, J. D., Dorff, C., Kenkel, B., & Larson, J. (2023) · doi:10.31235/osf.io/5ecfa · Cited by 213
van Es, K., & Nguyen, D. (2025) · AI & SOCIETY · doi:10.1007/s00146-024-02108-6 · Cited by 37
“…perpetuating existing biases and inequalities (Bender et al. 2021; Bianchi et al. 2022), exploiting workers in the development of models like GPT (Williams et al. 2022), and the significant environmental costs of training and maintaining these models (Bender et al. 2021; Luccioni et al. 2023).”
Weisman, D., Sugarman, A., Huang, Y. M., Gelberg, L., Ganz, P. A., & Comulada, W. S. (2025) · JMIR Formative Research · doi:10.2196/65670 · Cited by 37
Using ChatGPT and Other Large Language Model (LLM) Applications for Academic Paper Assignments
Jungherr, A. (2023) · doi:10.31235/osf.io/d84q6 · Cited by 37
AI and Blackness: Toward Moving Beyond Bias and Representation
Dancy, C. L., & Saucier, P. K. (2021) · IEEE Transactions on Technology and Society · doi:10.1109/tts.2021.3125998 · Cited by 52
“As discussed by [46], these data pulled from several sources often represent the “voices of people most likely to hew to a hegemonic viewpoint”.”
BackgroundInfluentialData-Centric and Model-Centric AI: Twin Drivers of Compact and Robust Industry 4.0 Solutions
Hamid, O. H. (2023) · Applied Sciences · doi:10.3390/app13052753 · Cited by 50
Lightweight transformers for clinical natural language processing
Rohanian, O., Nouriborji, M., Jauncey, H., Kouchaki, S., Nooralahzadeh, F., Clifton, L., Merson, L., & Clifton, D. A. (2024) · Natural Language Engineering · doi:10.1017/s1351324923000542 · Cited by 65
“…build these progressively larger and more resource-intensive models is a subject of debate in the Artificial Intelligence (AI) community (Bender et al. 2021), and there is interest in developing methods that would make these tools more accessible by creating smaller and faster versions…”
MethodMildner, T., Cooney, O., Meck, A.-M., Bartl, M., Savino, G.-L., Doyle, P. R., Garaialde, D., Clark, L., Sloan, J., Wenig, N., et al. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642542 · Cited by 53
“A lack of data representing marginalised social groups increases their marginalisation even further.”
Thakkar, H., & Manimaran, A. (2023) · 2023 International Conference on Emerging Research in Computational Science (ICERCS) · doi:10.1109/icercs57948.2023.10434081 · Cited by 42
“These features of focused enhancements, in combination with its attention mechanisms makes Mistral 7B , an efficient language model in both performance also well as resource structured in contrast to the traditional transformer architecture of LLaMa[12]. make Mistral 7B a high-performance and resource-efficient langua…”
BackgroundMarres, N., Castelle, M., Gobbo, B., Poletti, C., & Tripp, J. (2024) · Big Data & Society · doi:10.1177/20539517241255103 · Cited by 62
McCoy, L. G., Ng, F. Y. C., Sauer, C. M., Legaspi, K. E. Y., Jain, B., Gallifant, J., McClurkin, M., Hammond, A., Goode, D., Gichoya, J. W., et al. (2024) · BMC Medical Education · doi:10.1186/s12909-024-06048-z · Cited by 63
“There are also risks specific to the LLM architecture and the primacy of text [17].”
BackgroundInvestigating machine moral judgement through the Delphi experiment
Jiang, L., Hwang, J. D., Bhagavatula, C., Bras, R. J. L., Liang, J. T., Levine, S., Dodge, J., Sakaguchi, K., Forbes, M., Hessel, J., et al. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-024-00969-6 · Cited by 97
Artificial intelligence: Augmenting telehealth with large language models
Snoswell, C. L., Snoswell, A. J., Kelly, J. T., Caffery, L. J., & Smith, A. C. (2025) · Journal of Telemedicine and Telecare · doi:10.1177/1357633x231169055 · Cited by 37
Stokoe, E. H., Albert, S., Buschmeier, H., & Stommel, W. (2024) · Discourse & Communication · doi:10.1177/17504813241267118 · Cited by 51
“While developers and engineers of conversational technologies do most of their work ‘under the hood’, there are myriad insights from CA to draw upon.”
Mirowski, P., Love, J., Mathewson, K., & Mohamed, S. (2024) · The 2024 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3630106.3658993 · Cited by 37
Weisz, J. D., Kumar, S. V., Muller, M., Browne, K.-E., Goldberg, A., Heintze, K. E., & Bajpai, S. (2025) · Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706599.3706670 · Cited by 37
“LLMs have the potential to reproduce materials contained within their training data [7]; in the domain of software engineering, there is a risk of contaminating a code-base with copyrighted material or code subject to specific licenses (e.g. GPL).”
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
“To address the distinct challenges of working with language data, such as those summarized by Bender et al. (2021), other researchers have proposed specialized documentation for work in NLP.”
BackgroundSelf-citationFaithful Reasoning Using Large Language Models
Creswell, A., & Shanahan, M. (2022) · arXiv · doi:10.48550/arxiv.2208.14271 · Cited by 37
Bandy, J., & Vincent, N. (2021) · arXiv · doi:10.48550/arxiv.2105.05241 · Cited by 36
Few-Shot Semantic Parsing with Language Models Trained on Code
Shin, R., & Van Durme, B. (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.396 · Cited by 36
Rillig, M. C., Mansour, I., Hempel, S., Bi, M., König‐Ries, B., & Kasirzadeh, A. (2024) · Ecology Letters · doi:10.1111/ele.14397 · Cited by 40
(Un)making AI Magic: A Design Taxonomy
Lupetti, M. L., & Murray-Rust, D. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3641954 · Cited by 58
“Here, we work with the idea of enchantment : “the experience of being caught up and carried away”, where our attention is captured, we are “charmed and disturbed”, and our “background sense of order has flown out the door” [12].”
BackgroundHow Do Analysts Understand and Verify AI-Assisted Data Analyses?
Gu, K., Shang, R., Althoff, T., Wang, C., & Drucker, S. M. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642497 · Cited by 37
“1) [48, 79, 137] or make seemingly correct erroneous outputs [13, 47].”
Carik, B., Ping, K., Ding, X., & Rho, E. H. R. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3701194 · Cited by 64
Generative AI and its dilemmas: exploring AI from a translanguaging perspective
Jeon, J., Wei, L., Tai, K. W. H., & Lee, S. (2025) · Applied Linguistics · doi:10.1093/applin/amaf049 · Cited by 60
North Carolina State University, U. S. A., Anson, C., Straume, I., & University of Oslo, N. (2022) · Journal of Academic Writing · doi:10.18552/joaw.v12i1.820 · Cited by 36
“People experimenting with GPT-3 do find that the system can create bizarre, illogical, incoherent, or repetitive text because it does not really ‘think,’ a problem that has led some to characterize such systems as ‘stochastic parrots’—repeating what they have heard and adding random material (Bender et al., 2021).”
BackgroundUrman, A., & Makhortykh, M. (2023) · doi:10.31219/osf.io/q9v8f · Cited by 35
The Emerging Social Status of Generative AI: Vocabularies of AI Competence in Public Discourse
Bran, E., Rughiniş, C., Nadoleanu, G., & Flaherty, M. G. (2023) · 2023 24th International Conference on Control Systems and Computer Science (CSCS) · doi:10.1109/cscs59211.2023.00068 · Cited by 36
The technology triad: disruptive AI, regulatory gaps and value change
Hopster, J. K. G., & Maas, M. M. (2023) · AI and Ethics · doi:10.1007/s43681-023-00305-5 · Cited by 56
The next generation of machine learning for tracking adaptation texts
Sietsma, A. J., Ford, J. D., & Minx, J. C. (2023) · Nature Climate Change · doi:10.1038/s41558-023-01890-3 · Cited by 55
Generative Artificial Intelligence: 8 Critical Questions for Libraries
Bridges, L. M., McElroy, K., & Welhouse, Z. (2024) · Journal of Library Administration · doi:10.1080/01930826.2024.2292484 · Cited by 57
“Most, if not all, LLMs were built on data scraped, usually without permission, from across the internet (Chesterman, 2023; Bender et al., 2021).”
BackgroundMets, M., Karjus, A., Ibrus, I., & Schich, M. (2024) · PLoS ONE · doi:10.1371/journal.pone.0302380 · Cited by 62
Padiyath, A., Hou, X., Pang, A., Viramontes Vargas, D., Gu, X., Nelson-Fromm, T., Wu, Z., Guzdial, M., & Ericson, B. (2024) · Proceedings of the 2024 ACM Conference on International Computing Education Research - Volume 1 · doi:10.1145/3632620.3671098 · Cited by 35
“LLMs are a class of machine learning models that probabilistically generate natural language and code by learning statistical relationships from text documents [9].”
Normative Challenges of Risk Regulation of Artificial Intelligence
Orwat, C., Bareis, J., Folberth, A., Jahnel, J., & Wadephul, C. (2024) · NanoEthics · doi:10.1007/s11569-024-00454-9 · Cited by 74
“Generative AI entails risks to fundamental rights such as stereotypes and discrimination, deep fakes and disinformation, or privacy violations (e.g., [51–53]).”
Under the Influence: A Survey of Large Language Models in Fake News Detection
Kuntur, S., Wróblewska, A., Paprzycki, M., & Ganzha, M. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3471735 · Cited by 74
Emergent AI-assisted discourse: a case study of a second language writer authoring with ChatGPT
Jacob, S. R., Tate, T. P., & Warschauer, M. (2024) · Journal of China Computer-Assisted Language Learning · doi:10.1515/jccall-2024-0011 · Cited by 102
Advancing Retrieval-Augmented Generation with Inverted Question Matching for Enhanced QA Performance
Saha, B., Saha, U., & Malik, M. Z. (2024) · IEEE Access · doi:10.1109/access.2024.3513155 · Cited by 85
“Hallucination, on the other hand, refers to instances when the model produces outputs that are linguistically coherent but factually inaccurate or irrelevant to the input query [12].”
Cut the bullshit: why GenAI systems are neither collaborators nor tutors
Flenady, G., & Sparrow, R. (2025) · Teaching in Higher Education · doi:10.1080/13562517.2025.2497263 · Cited by 50
Hullman, J. R., Kapoor, S., Nanayakkara, P., Gelman, A., & Narayanan, A. (2022) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3514094.3534196 · Cited by 59
Stade, E. C., Stirman, S. W., Ungar, L., et al. (2023) · doi:10.31234/osf.io/cuzvr · Cited by 35
Ding, Z., Srinivasan, A., Macneil, S., & Chan, J. (2023) · Creativity and Cognition · doi:10.1145/3591196.3593516 · Cited by 35
“…previous analyses of the ethics of LLMs based not just on direct harm to the user but also participation in harmful and oppressive systems of power [2, 6], we consider the following primary categories of harm: 1) perpetuating hegemonic systems of oppression via uncritical incorporation of harmful…”
BackgroundMethodInfluentialAcademic Integrity and Artificial Intelligence: An Overview
Kumar, R., Eaton, S. E., Mindzak, M., & Morrison, R. (2023) · Handbook of Academic Integrity · doi:10.1007/978-981-287-079-7_153-1 · Cited by 35
Surveying Generative AI's Economic Expectations
Bybee, L. (2023) · arXiv · doi:10.48550/arxiv.2305.02823 · Cited by 35
Towards intellectual freedom in an AI Ethics Global Community
Ebell, C., Baeza-Yates, R., Benjamins, R., Cai, H., Coeckelbergh, M., Duarte, T., Hickok, M., Jacquet, A., Kim, A., Krijger, J., et al. (2021) · AI and Ethics · doi:10.1007/s43681-021-00052-5 · Cited by 34
An AI ethics ‘David and Goliath’: value conflicts between large tech companies and their employees
Ryan, M., Christodoulou, E., Antoniou, J., & Iordanou, K. (2022) · AI & Society · doi:10.1007/s00146-022-01430-1 · Cited by 49
“One of the factors that led to Gebru’s departure from Google was her critical analysis of using large language models (Bender et al. 2021).”
Lankford, S., Afli, H., & Way, A. (2023) · Information · doi:10.3390/info14120638 · Cited by 70
“Motivated by research which examines the environmental impact of NLP [3,49], we monitored the energy and carbon emissions required to train our models.”
MethodA Systematic Review of Fairness, Accountability, Transparency, and Ethics in Information Retrieval
Bernard, N., & Balog, K. (2025) · ACM Computing Surveys · doi:10.1145/3637211 · Cited by 67
Artificial Intelligence and Bias Towards Marginalised Groups: Theoretical Roots and Challenges
Mergen, A., Çetin-Kiliç, N., & Özbilgin, M. F. (2025) · International perspectives on equality, diversity and inclusion · doi:10.1108/s2051-233320250000012004 · Cited by 59
TOGLL: Correct and Strong Test Oracle Generation with LLMS
Hossain, S. B., & Dwyer, M. B. (2025) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1109/icse55347.2025.00098 · Cited by 110
“As fine-tuning and inference with large models is environmentally and financially costly [4], small models are most preferred.”
Eve, M. P. (2022) · OAPEN (The OAPEN Foundation) · Cited by 34
Language and culture internalization for human-like autotelic AI
Colas, C., Karch, T., Moulin-Frier, C., & Oudeyer, P.-Y. (2022) · Nature Machine Intelligence · doi:10.1038/s42256-022-00591-4 · Cited by 34
Using causal models to bridge the divide between big data and educational theory
Kitto, K., Hicks, B., & Shum, S. B. (2023) · British Journal of Educational Technology · doi:10.1111/bjet.13321 · Cited by 50
Surveying Generative AI's Economic Expectations
Bybee, L. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4430515 · Cited by 35
“4That LLMs reflect the biases and beliefs of their training corpus is central to much of the current work on AI alignment (Bender et al. (2021), Schramowski et al. (2022)) First, I validate how well GPT’s expectations of the stock market match the expectations of the American Association of…”
CounterFAccTual: How FAccT Undermines Its Organizing Principles
Gansky, B., & McDonald, S. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533241 · Cited by 34
“While this isn’t always the case, especially for the FAccT community - which has proven the environmental costs, for example, of building and training machine learning models ([6, 13]) - most of the governance frameworks focused on research and prototyping”
BackgroundTech Worker Organizing for Power and Accountability
Boag, W., Suresh, H., Lepe, B., & D'Ignazio, C. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533111 · Cited by 34
Ashkinaze, J., Mendelsohn, J., Qiwei, L., Budak, C., & Gilbert, E. (2025) · Proceedings of the ACM Collective Intelligence Conference · doi:10.1145/3715928.3737481 · Cited by 33
Translation Technology and Ethical Competence: An Analysis and Proposal for Translators’ Training
Polo, L. R., & Vargas-Sierra, C. (2023) · Languages · doi:10.3390/languages8020093 · Cited by 74
A Practical Guide to Conversation Research: How to Study What People Say to Each Other
Yeomans, M., Boland, F. K., Collins, H. K., Abi-Esber, N., & Brooks, A. W. (2023) · Advances in Methods and Practices in Psychological Science · doi:10.1177/25152459231183919 · Cited by 63
“However, transparency is necessary but not sufficient—many modern NLP models are still too complex to scrutinize, even by experts (Bender et al., 2021).”
BackgroundPeng, Y., Lock, I., & Salah, A. A. (2023) · Communication Methods and Measures · doi:10.1080/19312458.2023.2277956 · Cited by 50
“However, the development of LLMs is computationally costly and may pose negative environmental impacts (Bender et al., 2021).”
BackgroundThe promises of large language models for protein design and modeling
Valentini, G., Malchiodi, D., Gliozzo, J., Mesiti, M., Soto-Gomez, M., Cabri, A., Reese, J., Casiraghi, E., & Robinson, P. N. (2023) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2023.1304099 · Cited by 48
AI for social good and the corporate capture of global development
Iazzolino, G., & Stremlau, N. (2024) · Information Technology for Development · doi:10.1080/02681102.2023.2299351 · Cited by 48
“The overreliance on AI systems could indeed contribute to the reproduction of structural inequalities and injustice built into the data sets used to train predictive and generative models (Bender et al., 2021; Birhane & Prabhu, 2021; Eubanks, 2018; Holzmeyer, 2021).”
Metadata
- Authors
- Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S.
- Affiliations
- University of Washington (US)
- Black in AI, Palo Alto, CA, USA
- The Aether
- Venue
- Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency
- Publisher
- ACM
- Published
- 2021-03
- Pages
- 610-623
- OpenAlex
- W3133702157
- Cited by
- 6,988
- References
- 117
- Citations by year
- 2026: 2484 · 2025: 1887 · 2024: 1153 · 2023: 842 · 2022: 384 · 2021: 229
- FWCI
- 300.72
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- ACM Conference on Fairness, Accountability, and Transparency (FAccT) (publishedVersion, cc-by)
- Creighton Digital Repository (CDR) (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Keywords
- large language models, model size, AI risks, financial costs, data curation, dataset documentation, stakeholder value, BERT, GPT, fine-tuning
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.