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 301–400 of 6,931
Artificial intelligence in the information ecosystem: Affordances for everyday information seeking
Hirvonen, N., Jylhä, V., Lao, Y., & Larsson, S. (2024) · Journal of the Association for Information Science and Technology · doi:10.1002/asi.24860 · Cited by 107
“Recent scholarship on increased awareness of the energy and human resources that the large AI models demand in both training and deployment (see Bender et al., 2021; Crawford, 2021), emphasize the need to consider the sustainability of their use.”
Robo-writers: the rise and risks of language-generating AI
Hutson, M. (2021) · Nature · doi:10.1038/d41586-021-00530-0 · Cited by 141
Using generative artificial intelligence/ChatGPT for academic communication: Students' perspectives
Liu, Y., Park, J., & McMinn, S. (2024) · International Journal of Applied Linguistics · doi:10.1111/ijal.12574 · Cited by 106
Language Model Behavior: A Comprehensive Survey
Chang, T. A., & Bergen, B. K. (2024) · Computational Linguistics · doi:10.1162/coli_a_00492 · Cited by 105
Advancing AI in healthcare: A comprehensive review of best practices
Polevikov, S. (2023) · Clinica Chimica Acta · doi:10.1016/j.cca.2023.117519 · Cited by 142
History, development, and principles of large language models: an introductory survey
Wang, Z., Chu, Z., Doan, T. V., Ni, S., Yang, M., & Zhang, W. (2025) · AI and Ethics · doi:10.1007/s43681-024-00583-7 · Cited by 109
“These aspects disproportionately affect already disadvantaged and marginalized groups [15, 22, 162].”
Interactive Model Cards: A Human-Centered Approach to Model Documentation
Crisan, A., Drouhard, M., Vig, J., & Rajani, N. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533108 · Cited by 101
Emerging trends: A gentle introduction to fine-tuning
Church, K., Chen, Z., & Ma, Y. (2021) · Natural Language Engineering · doi:10.1017/s1351324921000322 · Cited by 165
Schrier, J., Norquist, A. J., Buonassisi, T., & Brgoch, J. (2023) · Journal of the American Chemical Society · doi:10.1021/jacs.3c04783 · Cited by 118
Responsible media technology and AI: challenges and research directions
Trattner, C., Jannach, D., Motta, E., Meijer, I. C., Diakopoulos, N. A., Elahi, M., Opdahl, A. L., Tessem, B., Borch, N. T., Fjeld, M., et al. (2021) · AI and Ethics · doi:10.1007/s43681-021-00126-4 · Cited by 166
Literacy in the Time of Artificial Intelligence
Kalantzis, M., & Cope, B. (2025) · Reading Research Quarterly · doi:10.1002/rrq.591 · Cited by 119
“They express bias to the extent that the texts they have captured express a history of bias (Bender et al., 2021; Magee et al., 2021).”
Rethinking machine unlearning for large language models
Liu, S., Yao, Y., Jia, J., Casper, S., Baracaldo, N., Hase, P., Yao, Y., Liu, C. Y., Xu, X., Li, H., et al. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-00985-0 · Cited by 643
“These include societal biases and stereotyping Bender et al. (2021); Motoki et al. (2023); Kotek et al. (2023), the generation of sensitive, private, harmful, or illegal content Nasr et al. (2023); Wen et al. (2023); Karamolegkou et al. (2023); Patil et al. (2024), ease of jailbreaking Wei et al.…”
Foundation Models: A New Paradigm for Artificial Intelligence
Schneider, J., Meske, C., & Kuss, P. (2024) · Business & Information Systems Engineering · doi:10.1007/s12599-024-00851-0 · Cited by 104
“This enabled the scaling of larger and more expressive pre-trained models that can be optimized in subsequent fine-tuning (Bender et al. 2021).”
InfluentialLarge Language Models Are Not Strong Abstract Reasoners
Gendron, G., Bao, Q., Witbrock, M., & Dobbie, G. (2023) · arXiv · doi:10.48550/arxiv.2305.19555 · Cited by 100
Exploring How Machine Learning Practitioners (Try To) Use Fairness Toolkits
Deng, W. H., Nagireddy, M., Lee, M. S. A., Singh, J., Wu, Z. S., Holstein, K., & Zhu, H. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533113 · Cited by 99
Designing Postdigital Futures: Which Designs? Whose Futures?
Macgilchrist, F., Allert, H., Pargman, T. C., & Jarke, J. (2023) · Postdigital Science and Education · doi:10.1007/s42438-022-00389-y · Cited by 138
Amirjalili, F., Neysani, M., & Nikbakht, A. (2024) · Frontiers in Education · doi:10.3389/feduc.2024.1347421 · Cited by 99
“Despite its ability to generate impressive outputs, the model may produce inaccuracies or nonsensical responses, demonstrating awareness of context and wording in input (Bender et al., 2021).”
Challenges and Strategies in Cross-Cultural NLP
Hershcovich, D., Frank, S., Lent, H., de Lhoneux, M., Abdou, M., Brandl, S., Bugliarello, E., Cabello Piqueras, L., Chalkidis, I., Cui, R., et al. (2022) · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2022.acl-long.482 · Cited by 99
Revisiting Weakly Supervised Pre-Training of Visual Perception Models
Singh, M., Gustafson, L., Adcock, A., De Freitas Reis, V., Gedik, B., Kosaraju, R. P., Mahajan, D., Girshick, R., Dollar, P., & Van Der Maaten, L. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00088 · Cited by 99
“A potential downside of weakly-supervised training of models on uncurated web data is that they may learn harmful associations that reflect offensive stereotypes [6, 9].”
BackgroundResultRoberts, J., Baker, M., & Andrew, J. (2024) · Critical Perspectives on Accounting · doi:10.1016/j.cpa.2024.102722 · Cited by 132
Evaluating the carbon footprint of NLP methods: a survey and analysis of existing tools
Bannour, N., Ghannay, S., Névéol, A., & Ligozat, A.-L. (2021) · Proceedings of the Second Workshop on Simple and Efficient Natural Language Processing · doi:10.18653/v1/2021.sustainlp-1.2 · Cited by 98
“The impact has been increasing over the years (Schwartz et al., 2019) and is affecting populations that can be different from those generating the impact (Bender et al., 2021).”
Backgroundvon der Mosel, J., Trautsch, A., & Herbold, S. (2022) · IEEE Transactions on Software Engineering · doi:10.1109/tse.2022.3178469 · Cited by 133
Reclaiming AI as a Theoretical Tool for Cognitive Science
van Rooij, I., Guest, O., Adolfi, F., de Haan, R., Kolokolova, A., & Rich, P. (2024) · Computational Brain & Behavior · doi:10.1007/s42113-024-00217-5 · Cited by 97
COLD: A Benchmark for Chinese Offensive Language Detection
Deng, J., Zhou, J., Sun, H., Zheng, C., Mi, F., Meng, H. M. L., & Huang, M. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.796 · Cited by 226
Care and Scale: Decorrelative Ethics in Algorithmic Recommendation
Seaver, N. (2021) · Cultural Anthropology · doi:10.14506/ca36.3.11 · Cited by 94
Generative AI meets Responsible AI: Practical Challenges and Opportunities
Kenthapadi, K., Lakkaraju, H., & Rajani, N. F. (2023) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3580305.3599557 · Cited by 158
“crawled from the internet, and consequently the models often reflect the practices of the wealthiest communities and countries [2].”
BackgroundJiang, P., Sonne, C., Li, W., You, F., & You, S. (2024) · Engineering · doi:10.1016/j.eng.2024.04.002 · Cited by 166
Contrasting Linguistic Patterns in Human and LLM-Generated News Text
Muñoz-Ortiz, A., Gómez-Rodríguez, C., & Vilares, D. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10903-2 · Cited by 94
“Most LLMs are trained totally or partially on scraped data from the web, and such unfiltered internet data usually contain biases or discrimination as they reproduce the hegemonic view (Bender et al. 2021).”
BackgroundSpennemann, D. H. R. (2023) · Knowledge · doi:10.3390/knowledge3030032 · Cited by 93
Wu, J., Cai, Y., Sun, T., Ma, K., & Lu, C. (2024) · Journal of Engineering Design · doi:10.1080/09544828.2024.2362587 · Cited by 134
“Bender et al. (2021) calls it a ‘Stochastic Parrot,’ which reproduces data used in the training process.”
Pretrained Transformers as Universal Computation Engines
Lu, K., Grover, A., Abbeel, P., & Mordatch, I. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.05247 · Cited by 91
“For high stakes applications in the real-world, there are potential concerns with transfer of harmful biases from one modality to one another using pretrained transformer models trained on vast quantities of unlabeled, uncurated datasets (Sheng et al., 2019; Bender et al., 2021).”
BackgroundWhen Large Language Models Meet Vector Databases: A Survey
Jing, Z., Su, Y., & Han, Y. (2025) · 2025 Conference on Artificial Intelligence x Multimedia (AIxMM) · doi:10.1109/aixmm62960.2025.00008 · Cited by 93
Nastasi, A. J., Courtright, K. R., Halpern, S. D., & Weissman, G. E. (2023) · Scientific Reports · doi:10.1038/s41598-023-45223-y · Cited by 93
Using Artificial Intelligence in TESOL: Some Ethical and Pedagogical Considerations
Pack, A., & Maloney, J. (2024) · TESOL Quarterly · doi:10.1002/tesq.3320 · Cited by 92
A phenomenology and epistemology of large language models: transparency, trust, and trustworthiness
Heersmink, R., de Rooij, B., Clavel Vázquez, M. J., & Colombo, M. (2024) · Ethics and Information Technology · doi:10.1007/s10676-024-09777-3 · Cited by 91
“For example, Bender et al. (2021) argue that LLMs pick on biases present in the data they were trained on.”
Cultural cartography with word embeddings
Stoltz, D. S., & Taylor, M. A. (2021) · Poetics · doi:10.1016/j.poetic.2021.101567 · Cited by 120
Why we need to be careful with LLMs in medicine
Bélisle-Pipon, J.-C. (2024) · Frontiers in Medicine · doi:10.3389/fmed.2024.1495582 · Cited by 91
Generative language models exhibit social identity biases
Hu, T., Kyrychenko, Y., Rathje, S., Collier, N., van der Linden, S., & Roozenbeek, J. (2024) · Nature Computational Science · doi:10.1038/s43588-024-00741-1 · Cited by 100
“GPT-2-Large Positive .5106 .9375 2 Results”
Designing Responsible AI: Adaptations of UX Practice to Meet Responsible AI Challenges
Wang, Q., Madaio, M., Kane, S., Kapania, S., Terry, M., & Wilcox, L. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581278 · Cited by 90
“Substantial prior work has identifed the potential harms of technologies built on LLMs [e.g., 8, 15, 105], while at the same time acknowledging the challenge of evaluating and mitigating such harms [14, 100, 123] and calling for HCI and UX researchers and practitioners to contribute a…”
BackgroundGenerative artificial intelligence as an enabler of student feedback engagement: a framework
Zhan, Y., Boud, D., Dawson, P., & Yan, Z. (2025) · Higher Education Research & Development · doi:10.1080/07294360.2025.2476513 · Cited by 94
“GenAI has been trained on datasets that dominantly reflect US standardised English conventions (Bender et al., 2021).”
Process for Adapting Language Models to Society (PALMS) with Values-Targeted Datasets
Solaiman, I., & Dennison, C. (2021) · arXiv · doi:10.48550/arxiv.2106.10328 · Cited by 90
“[4] describe GPT systems encoding harmful bias across identities, including abusive language patterns.”
BackgroundMind the Gap: Assessing Temporal Generalization in Neural Language Models
Lazaridou, A., Kuncoro, A., Gribovskaya, E., Agrawal, D., Liska, A., Terzi, T., Gimenez, M., d'Autume, C. D. M., Kocisky, T., Ruder, S., et al. (2021) · arXiv · doi:10.48550/arxiv.2102.01951 · Cited by 90
“For instance, recently there has been renewed public support and interest for social justice movements in 2020, such as the #BlackLivesMatter movement (Cohn and Quealy, 2020).”
BackgroundInfluentialWhat Should ChatGPT Mean for Bioethics?
Cohen, I. G. (2023) · The American Journal of Bioethics · doi:10.1080/15265161.2023.2233357 · Cited by 98
“…way in which one might worry that the models are not big enough — the sampling thus far has focused on English language sources, but of course users hail from many different language groups, and the searchable internet represents only a particular slice of human communication (Bender et al. 2021).”
BackgroundWhat do LLMs need to Synthesize Correct Router Configurations?
Mondal, R., Tang, A., Beckett, R., Millstein, T., & Varghese, G. (2023) · Proceedings of the 22nd ACM Workshop on Hot Topics in Networks · doi:10.1145/3626111.3628194 · Cited by 90
Large language models for preventing medication direction errors in online pharmacies
Pais, C., Liu, J., Voigt, R., Gupta, V., Wade, E., & Bayati, M. (2024) · Nature Medicine · doi:10.1038/s41591-024-02933-8 · Cited by 89
“…and metric utilized to assess the performance of each AI approach. e , Data description and how the data were used in the study to train and evaluate the different AI approaches. https://doi.org/10.1038/s41591-024-02933-8 decision support systems to enhance patient safety 25,26 , this…”
InfluentialLLM Fine-Tuning: Concepts, Opportunities, and Challenges
Wu, X.-K., Chen, M., Li, W., Wang, R., Lu, L., Liu, J., Hwang, K., Hao, Y., Pan, Y., Meng, Q., et al. (2025) · Big Data and Cognitive Computing · doi:10.3390/bdcc9040087 · Cited by 158
“In the context of human–AI interaction, this “distance” signifies the cognitive gap between human understanding and machine interpretation, highlighting the challenges of bridging the humans and AI comprehension [17].”
Embedded reproduction in platform data work
Posada, J. (2022) · Information, Communication & Society · doi:10.1080/1369118x.2022.2049849 · Cited by 89
“mented to produce environmental damage to a certain extent (Bender et al., 2021; Dauvergne, 2020), my claim here is not that data work has a substantial impact on the environment, but rather that by being embedded in nature, the harms derived from pollution and other forms of environmental damage affect the livelihood…”
BackgroundCheong, I., Xia, K., Feng, K. J. K., Chen, Q. Z., & Zhang, A. X. (2024) · The 2024 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3630106.3659048 · Cited by 91
The Ethics of AI Ethics. A Constructive Critique
Heilinger, J. (2022) · Philosophy & Technology · doi:10.1007/s13347-022-00557-9 · Cited by 145
“…these systems is increasingly receiving attention in the AI-ethical literature, e.g. with regard to the energy needed for training neural networks (Bender et al., 2021; Strubell et al., 2019); when showing that AI-based technologies cannot only be used to develop strategies for sustainable…”
BackgroundIs neuro-symbolic AI meeting its promises in natural language processing? A structured review
Hamilton, K., Nayak, A., Božić, B., & Longo, L. (2024) · Semantic Web · doi:10.3233/sw-223228 · Cited by 88
Language-based game theory in the age of artificial intelligence
Capraro, V., Paolo, R. D., Perc, M., & Pizziol, V. (2024) · Journal of The Royal Society Interface · doi:10.1098/rsif.2023.0720 · Cited by 115
Exploring the use of artificial intelligence (AI) in the delivery of effective feedback
Venter, J., Coetzee, S. A., & Schmulian, A. (2024) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2024.2415649 · Cited by 156
“LLMs create responses by predicting the likelihood of word sequences based on large amounts of data (Bender et al. 2021; Marcus and Davis 2020).”
Possible harms of Artificial Intelligence and the EU AI Act: fundamental rights and risk
Kusche, I. (2024) · Journal of Risk Research · doi:10.1080/13669877.2024.2350720 · Cited by 87
Klein, L. F., & D’Ignazio, C. (2024) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3630106.3658543 · Cited by 173
People Overtrust AI-Generated Medical Advice despite Low Accuracy
Shekar, S., Pataranutaporn, P., Sarabu, C. R., Cecchi, G., & Maes, P. (2025) · NEJM AI · doi:10.1056/aioa2300015 · Cited by 139
“56,57 The stakes are high in medical applications.”
At the Tensions of South and North: Critical Roles of Global South Stakeholders in AI Governance
Png, M.-T. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533200 · Cited by 85
AI’s Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia
Qadri, R., Shelby, R., Bennett, C. L., & Denton, R. (2023) · 2023 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3593013.3594016 · Cited by 85
When Do You Need Billions of Words of Pretraining Data?
Zhang, Y., Warstadt, A., Li, X., & Bowman, S. R. (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.90 · Cited by 85
Why ‘open’ AI systems are actually closed, and why this matters
Widder, D. G., Whittaker, M., & West, S. M. (2024) · Nature · doi:10.1038/s41586-024-08141-1 · Cited by 86
Identifying and Improving Disability Bias in GPT-Based Resume Screening
Glazko, K., Mohammed, Y., Kosa, B., Potluri, V., & Mankoff, J. (2024) · The 2024 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3630106.3658933 · Cited by 85
“Since “... society’s racism, misogyny, ableism, etc., tend to be overrep-resented in training data ... [an LLM] that has been trained on such data will pick up these kinds of problematic associations” [4].”
BackgroundA 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
“Some authors have called for smaller reusable models that are trained on contextually relevant, rather than large, datasets [10].”
BackgroundImplicit Representations of Meaning in Neural Language Models
Li, B. Z., Nye, M., & Andreas, 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.143 · Cited by 84
Applying Large Language Models to Power Systems: Potential Security Threats
Ruan, J., Liang, G., Zhao, H. Q., Liu, G., Sun, X., Qiu, J., Xu, Z., Wen, F., & Dong, Z. Y. (2024) · IEEE Transactions on Smart Grid · doi:10.1109/tsg.2024.3373256 · Cited by 104
“As ultra-large-scale neural networks, LLMs require substantial computational resources and training investment to ensure performance [9].”
BackgroundCharacterizing Power Management Opportunities for LLMs in the Cloud
Patel, P., Choukse, E., Zhang, C., Goiri, Í., Warrier, B., Mahalingam, N., & Bianchini, R. (2024) · Proceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3 · doi:10.1145/3620666.3651329 · Cited by 100
“Cloud providers and datacenter operators today face a massive GPU capacity crunch due to the explosion in demand for large language models (LLMs) [8].”
BackgroundMirza, A., Alampara, N., Kunchapu, S., Ríos-García, M., Emoekabu, B., Krishnan, A., Gupta, T., Schilling-Wilhelmi, M., Okereke, M., Aneesh, A., et al. (2025) · Nature Chemistry · doi:10.1038/s41557-025-01815-x · Cited by 380
The AI Assessment Scale (AIAS) in action: A pilot implementation of GenAI supported assessment
Furze, L., Perkins, M., Roe, J., & MacVaugh, J. (2024) · Australasian Journal of Educational Technology · doi:10.14742/ajet.9434 · Cited by 92
“The datasets used to train these models, such as ImageNet and The Pile, often reflect the biases and worldviews of the predominantly white, middle class, and English-speaking males who contribute to them (Bender et al., 2021; Crawford, 2021).”
Haber, Y., Levkovich, I., Hadar-Shoval, D., & Elyoseph, Z. (2024) · JMIR Mental Health · doi:10.2196/54781 · Cited by 81
“While GAI systems have demonstrated impressive language processing and generation capabilities, they still do not possess genuine understanding or comprehension of language in the same way that humans do [17].”
BackgroundTrustworthy AI: Securing Sensitive Data in Large Language Models
Feretzakis, G., & Verykios, V. S. (2024) · AI · doi:10.3390/ai5040134 · Cited by 83
“It gives very little insight into whether one can predict or control what the model may reveal [20].”
Mirages. On Anthropomorphism in Dialogue Systems
Abercrombie, G., Curry, A., Dinkar, T., Rieser, V., & Talat, Z. (2023) · Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2023.emnlp-main.290 · Cited by 83
“…of Anthropomor-phic Tools Given our inherent nature to attribute meaning to signals, one must consider the appropriateness of the tool and use cases (Bender et al., 2021; Dinan et al., 2022) when designing dialogue systems, in order to avoid the (over-)integration of anthropomorphic cues.”
MethodArtificial intelligence and the climate emergency: Opportunities, challenges, and recommendations
Taddeo, M., Tsamados, A., Cowls, J., & Floridi, L. (2021) · One Earth · doi:10.1016/j.oneear.2021.05.018 · Cited by 113
Making Sense of the Digital Automation of Education
Selwyn, N., Hillman, T., Rensfeldt, A. B., & Perrotta, C. (2022) · Postdigital Science and Education · doi:10.1007/s42438-022-00362-9 · Cited by 85
“As implied in recent critiques of automated language systems as ‘stochastic parrots’ (Bender et al. 2021), we automate complex human processes such as sense-making and meaning-making at our own risk.”
BackgroundChatting with the dead: The hermeneutics of thanabots
Henrickson, L. (2023) · Media Culture & Society · doi:10.1177/01634437221147626 · Cited by 112
GREENER principles for environmentally sustainable computational science
Lannelongue, L., Aronson, H. G., Bateman, A. G., Birney, E., Caplan, T., Juckes, M. N., McEntyre, J., Morris, A. D., Reilly, G., & Inouye, M. (2023) · Nature Computational Science · doi:10.1038/s43588-023-00461-y · Cited by 116
Yildirim, N., Richardson, H., Wetscherek, M. T., Bajwa, J., Jacob, J., Pinnock, M. A., Harris, S., Coelho De Castro, D., Bannur, S., Hyland, S., et al. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642013 · Cited by 82
“…include questions around domain specificity and quality of model input data [47, 74] as well as societal biases that are inherent in that data [12], which increase with model scale and multimodality [18], risking harms by exacerbating health disparities and social inequalities [11, 118, 125, 131].”
BackgroundDoes ChatGPT have semantic understanding? A problem with the statistics-of-occurrence strategy
Titus, L. M. (2024) · Cognitive Systems Research · doi:10.1016/j.cogsys.2023.101174 · Cited by 82
“See Bender et al. (2021), for example, who assess language models as mere stochastic parrots given our understanding of how they are trained and how they function.”
Prabhudesai, S., Yang, L., Asthana, S., Huan, X., Liao, Q. V., & Banovic, N. (2023) · Proceedings of the 28th International Conference on Intelligent User Interfaces · doi:10.1145/3581641.3584033 · Cited by 82
“Additionally, suppressing uncertainty grossly oversimplifies the complex, dynamic and ambiguous nature of the world [42], which can lead to over-reliance of decision-makers on inaccurate DSS predictions that may be rooted in harmful historical biases [2, 10, 12, 35, 81, 82].”
BackgroundLuitse, D. M. R. (2024) · Internet Policy Review · doi:10.14763/2024.2.1768 · Cited by 124
Entrepreneurship Education at the Dawn of Generative Artificial Intelligence
Winkler, C., Hammoda, B., Noyes, E., & Van Gelderen, M. (2023) · Entrepreneurship Education and Pedagogy · doi:10.1177/25151274231198799 · Cited by 81
A Survey on Trustworthy Recommender Systems
Ge, Y., Liu, S., Fu, Z., Tan, J., Li, Z., Xu, S., Li, Y., Xian, Y., & Zhang, Y. (2024) · ACM Transactions on Recommender Systems · doi:10.1145/3652891 · Cited by 235
An Autoethnographic Case Study of Generative Artificial Intelligence's Utility for Accessibility
Glazko, K. S., Yamagami, M., Desai, A., Mack, K. A., Potluri, V., Xu, X., & Mankoff, J. (2023) · The 25th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3597638.3614548 · Cited by 80
“Un-surprisingly, some of the limitations we encountered reflect GAI’s current ability to parrot [5] information without truly integrating it.”
BackgroundAbundant intelligences: placing AI within Indigenous knowledge frameworks
Lewis, J. E., Whaanga, H., & Yolgörmez, C. (2025) · AI & SOCIETY · doi:10.1007/s00146-024-02099-4 · Cited by 80
“For example, the environmental cost of AI systems has been a significant point of contention, especially generative AI technologies, but also others, because of their resource-intensiveness (Bender et al. 2021; Strubell et al. 2019).”
The Future of Research in an Artificial Intelligence-Driven World
Kulkarni, M., Mantere, S., Vaara, E., van den Broek, E., Pachidi, S., Glaser, V. L., Gehman, J., Petriglieri, G., Lindebaum, D., Cameron, L. D., et al. (2024) · Journal of Management Inquiry · doi:10.1177/10564926231219622 · Cited by 80
ChatGPT: deconstructing the debate and moving it forward
Coeckelbergh, M., & Gunkel, D. J. (2024) · AI & SOCIETY · doi:10.1007/s00146-023-01710-4 · Cited by 78
“A similar statement has been provided by Emily Bender—a linguist and co-author of the “Stochostic Parrots” essay (Bender et al. 2021) that discusses some of the risks associated with large language models—in a recent profile that was published in New York Magazine : “How should we interpret the…”
Hudley, A. H. C., Mallinson, C., & Bucholtz, M. (2024) · doi:10.1093/oso/9780197755303.001.0001 · Cited by 94
Friend or foe? Exploring the implications of large language models on the science system
Fecher, B., Hebing, M., Laufer, M., Pohle, J., & Sofsky, F. (2025) · AI & SOCIETY · doi:10.1007/s00146-023-01791-1 · Cited by 78
“This is considered a structural issue of how these systems are trained and cannot be resolved by simply creating bigger models as size does not guarantee diversity (Bender et al. 2021).”
The 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
“AI systems can propagate harms such as exacerbating social inequity [64, 7, 41] and harmful biases [20, 13], which can be further amplified in larger systems as scale increases [6].”
BackgroundMarkl, N. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533117 · Cited by 78
On the Design of AI-powered Code Assistants for Notebooks
Mcnutt, A. M., Wang, C., Deline, R. A., & Drucker, S. M. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3580940 · Cited by 77
Do Language Models’ Words Refer?
Mandelkern, M., & Linzen, T. (2024) · Computational Linguistics · doi:10.1162/coli_a_00522 · Cited by 77
“But some have argued that LMs are really nothing more than babbling “stochastic parrots” (a term created by Bender et al. 2021).”
BackgroundJang, S., Lee, G., Oh, J., Lee, J., & Koo, B. (2024) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.102532 · Cited by 77
The power of emotions: Leveraging user generated content for customer experience management
Sykora, M. D., Elayan, S., Hodgkinson, I. R., Jackson, T., & West, A. A. (2022) · Journal of Business Research · doi:10.1016/j.jbusres.2022.02.048 · Cited by 120
Choi, D., Lee, S., Kim, S.-I., Lee, K., Yoo, H. J., Lee, S., & Hong, H. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3641989 · Cited by 78
“One of the most active areas of research is the harmful dialogue produced by LLMs, which often originates from biased data or inadequate data representation [7, 10, 18, 30, 35, 37, 107, 116].”
BackgroundCultural Bias in Large Language Models: A Comprehensive Analysis and Mitigation Strategies
Liu, Z. (2025) · Journal of Transcultural Communication · doi:10.1515/jtc-2023-0019 · Cited by 77
“Typically trained on vast amounts of uncurated Internet-based data, LLMs inherit stereotypes, misrepresentations, derogatory language, and other harmful behaviors that disproportionately a ff ect already-vulnerable and marginalized communities (Bender et al. 2021; Dodge et al. 2021).”
Frontier AI Regulation: Managing Emerging Risks to Public Safety
Anderljung, M., Barnhart, J., Korinek, A., Leung, J., O'Keefe, C., Whittlestone, J., Avin, S., Brundage, M., Bullock, J., Cass-Beggs, D., et al. (2023) · arXiv · doi:10.48550/arxiv.2307.03718 · Cited by 77
The Sociodemographic Biases in Machine Learning Algorithms: A Biomedical Informatics Perspective
Franklin, G., Stephens, R., Piracha, M., Tiosano, S., Lehouillier, F., Koppel, R., & Elkin, P. (2024) · Life · doi:10.3390/life14060652 · Cited by 76
The carbon emissions of writing and illustrating are lower for AI than for humans
Tomlinson, B., Black, R. W., Patterson, D. J., & Torrance, A. W. (2024) · Scientific Reports · doi:10.1038/s41598-024-54271-x · Cited by 77
“Many relevant issues have been discussed elsewhere 29,30 ; here, we touch on a few items of most relevance to AI for writing and illustrating.”
BackgroundCyBERT: Cybersecurity Claim Classification by Fine-Tuning the BERT Language Model
Ameri, K., Hempel, M., Sharif, H., Lopez Jr., J., & Perumalla, K. (2021) · Journal of Cybersecurity and Privacy · doi:10.3390/jcp1040031 · Cited by 76
“Bias in machine learning often is the result of bias present in the training data set and is a known problem for any language model [35], including when training CyBERT.”
BackgroundCalling BS on ChatGPT: Reflections on AI as a Communication Source
Sundar, S. S., & Liao, M. (2023) · Journalism & Communication Monographs · doi:10.1177/15226379231167135 · Cited by 114
“Perhaps, the best term to describe ChatGPT is “stochastic parrot”: it generates content based on statistical and probabilistic information, without actual reference to meanings (Bender et al., 2021).”
BackgroundEnsemble Large Language Models: A Survey
Mienye, I. D., & Swart, T. G. (2025) · Information · doi:10.3390/info16080688 · Cited by 111
Epistemic Injustice in Generative AI
Kay, J., Kasirzadeh, A., & Mohamed, S. (2024) · Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1609/aies.v7i1.31671 · Cited by 76
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.