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 601–700 of 6,931
A Survey of AI Text-to-Image and AI Text-to-Video Generators
Singh, A. (2023) · International Conference on Artificial Intelligence, Robotics and Control (AIRC) · doi:10.1109/airc57904.2023.10303174 · Cited by 100
Chain-of-Thought Reasoning in Tabular Language Models
Zheng, M., Yang, H., Jiang, W., Lin, Z., Lyu, Y., She, Q., & Wang, W. (2023) · Findings of the Association for Computational Linguistics: EMNLP 2023 · doi:10.18653/v1/2023.findings-emnlp.734 · Cited by 50
“…instance, it is financially expensive to satisfy the high computational requirements, the storage capacity and the desired band-width of LLMs, which makes it a challenge for individual users or small organizations to utilize LLMs in their applications (Strubell et al., 2019; Bender et al., 2021).”
BackgroundStrong and weak AI narratives: an analytical framework
Bory, P., Natale, S., & Katzenbach, C. (2024) · AI & Society · doi:10.1007/s00146-024-02087-8 · Cited by 106
“However, the fact that ChatGPT remains, as Bender et al. (2021) famously put it, a “stochastic parrot” whose functioning and capacity cannot be compared to human intelligence and sentience, would amount to a failure only if one assumes that it needs to be assessed against the threshold of strong AI.”
AirGPT: Pioneering the convergence of conversational AI with atmospheric science
Song, J., Ma, C., & Ran, M. (2025) · npj Climate and Atmospheric Science · doi:10.1038/s41612-025-01070-4 · Cited by 48
Alvero, A. J., Lee, J., Regla-Vargas, A., Kizilcec, R. F., Joachims, T., & Antonio, A. L. (2024) · Journal of Big Data · doi:10.1186/s40537-024-00986-7 · Cited by 48
“A better understanding of LLMs will therefore require consideration of the bias literature alongside studies of sociolinguistic profiles of LLMs. Extensive research has uncovered notable biases within LLMs [13, 25, 32, 47, 50, 63, 65, 80].”
BackgroundCoelho Mollo, D., & Millière, R. (2026) · Philosophy and the Mind Sciences · doi:10.33735/phimisci.2026.12307 · Cited by 48
Quantifying Societal Bias Amplification in Image Captioning
Hirota, Y., Nakashima, Y., & García, N. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01309 · Cited by 92
Hate Speech Classifiers Learn Normative Social Stereotypes
Davani, A. M., Atari, M., Kennedy, B., & Dehghani, M. (2023) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00550 · Cited by 133
“Here, we rely on semantic representations of social groups in pre-trained language models, known to encode normative social stereotypes and biases of large text corpora (Bender et al., 2021).”
MethodHow AI tools can—and cannot—help organizations become more ethical
De Cremer, D., & Narayanan, D. (2023) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2023.1093712 · Cited by 48
Resisting Dehumanization in the Age of “AI”
Bender, E. M. (2024) · Current Directions in Psychological Science · doi:10.1177/09637214231217286 · Cited by 97
Self-citationSTELA: a community-centred approach to norm elicitation for AI alignment
Bergman, S., Marchal, N., Mellor, J., Mohamed, S., Gabriel, I., & Isaac, W. (2024) · Scientific Reports · doi:10.1038/s41598-024-56648-4 · Cited by 48
Reimagining Higher Education: Navigating the Challenges of Generative AI Adoption
Hughes, L., Malik, T., Dettmer, S., Al-Busaidi, A. S., & Dwivedi, Y. K. (2025) · Information Systems Frontiers · doi:10.1007/s10796-025-10582-6 · Cited by 125
Large Language Models for Text Classification: From Zero-Shot Learning to Instruction-Tuning
Chae, Y., & Davidson, T. (2025) · Sociological Methods & Research · doi:10.1177/00491241251325243 · Cited by 103
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 103
PIGLeT: Language Grounding Through Neuro-Symbolic Interaction in a 3D World
Zellers, R., Holtzman, A., Peters, M., Mottaghi, R., Kembhavi, A., Farhadi, A., & Choi, Y. (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.159 · Cited by 48
“However, this has unintended consequences (Bender et al., 2021) and it is unlike how children learn language, with children learning novel words from experience (Carey and Bartlett, 1978).”
BackgroundCan language models automate data wrangling?
Jaimovitch-López, G., Ferri, C., Hernández-Orallo, J., Martínez-Plumed, F., & Ramírez-Quintana, M. J. (2023) · Machine Learning · doi:10.1007/s10994-022-06259-9 · Cited by 48
“It is then very important to determine whether language models could be used in the future for data wrangling tasks, and whether they get better as the number of parameters increase, a question subject to recent debate (Bender et al., 2021; Tamkin et al., 2021).”
BackgroundTowards Measuring the Representation of Subjective Global Opinions in Language Models
Durmus, E., Nguyen, K., Liao, T. I., Schiefer, N., Askell, A., Bakhtin, A., Chen, C., Hatfield-Dodds, Z., Hernandez, D., Joseph, N., et al. (2023) · arXiv · doi:10.48550/arxiv.2306.16388 · Cited by 48
Hagendorff, T. (2021) · Minds and Machines · doi:10.1007/s11023-021-09573-8 · Cited by 47
Ethics of Socially Disruptive Technologies: An Introduction
van de Poel, I., Frank, L. E., Hermann, J., Hopster, J. K. G., Lenzi, D., Nyholm, S. R., Taebi, B., & Ziliotti, E. (2023) · Open Book Publishers · doi:10.11647/obp.0366 · Cited by 65
Towards Feature Selection for Ranking and Classification Exploiting Quantum Annealers
Dacrema, M. F., Moroni, F., Nembrini, R., Ferro, N., Faggioli, G., & Cremonesi, P. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531755 · Cited by 66
Chin, H., Lima, G., Shin, M., Zhunis, A., Cha, C., Choi, J., & Cha, M. (2023) · Journal of Medical Internet Research · doi:10.2196/40922 · Cited by 49
A Systematic Investigation of Commonsense Knowledge in Large Language Models
Li, X. L., Kuncoro, A., Hoffmann, J., d’Autume, C. D. M., Blunsom, P., & Nematzadeh, A. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.812 · Cited by 133
“Furthermore, the enormous size of the training data does not necessarily guarantee diversity: We expect the training data to contain a smaller proportion of vernacular or regional English that is used by underrepresented communities (Blodgett et al., 2016; Bender et al., 2021).”
BackgroundYong-In University, & Van Horn, K. (2024) · Teaching English as a Second or Foreign Language--TESL-EJ · doi:10.55593/ej.28109a8 · Cited by 47
Tung, J. Y. M., Gill, S. R., Sng, G. G. R., Lim, D. Y. Z., Ke, Y., Tan, T. F., Jin, L., Elangovan, K., Ong, J. C. L., Abdullah, H. R., et al. (2024) · Journal of Medical Internet Research · doi:10.2196/57721 · Cited by 59
“To reduce the risk of hallucination, prompts for safety included specific instructions to include only information provided in the fictional EMR [12].”
Fang, L., Yu, X., Cai, J., Chen, Y., Wu, S., Liu, Z., Yang, Z., Lu, H., Gong, X., Liu, Y., et al. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11423-3 · Cited by 128
What is it like to program with artificial intelligence?
Sarkar, A., Gordon, A. D., Negreanu, C., Poelitz, C., Ragavan, S. S., & Zorn, B. (2022) · arXiv · doi:10.48550/arxiv.2208.06213 · Cited by 47
“Provenance, too is lost: it is difficult to determine whether the generation is original to the model, or a stochastic parroting (Bender et al., 2021; Ziegler, 2021).”
BackgroundMollo, D. C., & Millière, R. (2023) · arXiv · doi:10.48550/arxiv.2304.01481 · Cited by 47
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy
Aujih, A. B., Shapiai, M. I., Mériaudeau, F., & Tang, T. B. (2022) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2022.3182907 · Cited by 51
Academic Integrity and Artificial Intelligence: An Overview
Kumar, R., Eaton, S. E., Mindzak, M., & Morrison, R. (2024) · Springer International Handbooks of Education · doi:10.1007/978-3-031-54144-5_153 · Cited by 46
Distilled GPT for source code summarization
Su, C.-Y., & McMillan, C. (2024) · Automated Software Engineering · doi:10.1007/s10515-024-00421-4 · Cited by 49
“); Bender et al (2021) introduce the technique to fine-tune Large Language Models (LLM) for code summarization.”
Computers as Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially
Alberts, L., Lyngs, U., & Kleek, M. G. V. (2024) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3653693 · Cited by 94
On the limitations of large language models in clinical diagnosis
Reese, J. T., Danis, D., Caufield, J. H., Groza, T., Casiraghi, E., Valentini, G., Mungall, C. J., & Robinson, P. N. (2023) · medRxiv · doi:10.1101/2023.07.13.23292613 · Cited by 46
How to design an AI ethics board
Schuett, J., Reuel, A.-K., & Carlier, A. (2025) · AI and Ethics · doi:10.1007/s43681-023-00409-y · Cited by 46
Large language models: an overview of foundational architectures, recent trends, and a new taxonomy
Mienye, I. D., Jere, N., Obaido, G., Ogunruku, O. O., Esenogho, E., & Modisane, C. (2025) · Discover Applied Sciences · doi:10.1007/s42452-025-07668-w · Cited by 46
“This can result in outputs that perpetuate harmful stereotypes or exhibit unfair treatment in contexts, such as job recruitment or legal decisions [159].”
Structured like a language model: Analysing AI as an automated subject
Magee, L., Arora, V., & Munn, L. (2023) · Big Data & Society · doi:10.1177/20539517231210273 · Cited by 46
The 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
Language Models Represent Space and Time
Gurnee, W., & Tegmark, M. (2023) · arXiv · doi:10.48550/arxiv.2310.02207 · Cited by 46
“Better understanding of if and how LLMs model the world is critical for reasoning about the robustness, fairness, and safety of current and future AI systems (Bender et al., 2021; Weidinger et al., 2022; Bommasani et al., 2021; Hendrycks et al., 2023; Ngo et al., 2023).”
BackgroundData 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
Generative AI Considered Harmful
Fischer, J. E. (2023) · Proceedings of the 5th International Conference on Conversational User Interfaces · doi:10.1145/3571884.3603756 · Cited by 47
Carbon Footprint of Selecting and Training Deep Learning Models for Medical Image Analysis
Selvan, R., Bhagwat, N., Anthony, L. F. W., Kanding, B., & Dam, E. B. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-16443-9_49 · Cited by 73
“In natural language processing (NLP), where some of the largest DL models are in use, the research community has taken notice of the growing energy consumption and its adverse impact on environmental sustainability and equitable research[31,4].”
BackgroundGeoAI, counter-AI, and human geography: A conversation
Janowicz, K., Sieber, R. E., & Crampton, J. W. (2022) · Dialogues in Human Geography · doi:10.1177/20438206221132510 · Cited by 45
“One of the most in fl uential papers that was presented at the FAccT (fairness, accountability, and transparency) conference, the same conference at which Bender et al. (2021) presented the famous Stochastic Parrots article, there was another paper on Chinese transfer learning and language models.”
BackgroundBushuven, S., Bentele, M., Bentele, S., Gerber, B., Bansbach, J., Ganter, J., Trifunovic-Koenig, M., & Ranisch, R. (2023) · Journal of Medical Systems · doi:10.1007/s10916-023-02019-x · Cited by 47
Copilots for Linguists: AI, Constructions, and Frames
Torrent, T. T., Hoffmann, T., Lorenzi, A., & Turner, M. B. (2023) · Cambridge University Press eBooks · doi:10.1017/9781009439190 · Cited by 62
Lee, M. H. J., Montgomery, J., & Lai, C. K. (2024) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3630106.3658975 · Cited by 106
‘AI coaching’: democratising coaching service or offering an ersatz?
Bachkirova, T., & Kemp, R. (2024) · Coaching An International Journal of Theory Research and Practice · doi:10.1080/17521882.2024.2368598 · Cited by 74
“One useful way of understanding LLMs is as ‘stochastic parrot’, a term coined by Bender et al. (2021).”
Digital Doppelgängers and Lifespan Extension: What Matters?
Iglesias, S., Earp, B. D., Voinea, C., Mann, S. P., Zahiu, A., Jecker, N. S., & Savulescu, J. (2024) · The American Journal of Bioethics · doi:10.1080/15265161.2024.2416133 · Cited by 63
Nastasi, A. J., Courtright, K. R., Halpern, S. D., & Weissman, G. E. (2023) · medRxiv · doi:10.1101/2023.02.25.23286451 · Cited by 44
Scarecrows in Oz: The Use of Large Language Models in HRI
Williams, T., Matuszek, C., Mead, R., & DePalma, N. (2024) · ACM Transactions on Human-Robot Interaction · doi:10.1145/3606261 · Cited by 67
Kenton, Z., Everitt, T., Weidinger, L., Gabriel, I., Mikulik, V., & Irving, G. (2021) · arXiv · doi:10.48550/arxiv.2103.14659 · Cited by 45
“Firstly, as mentioned in the introduction, we have recently seen impressive advances in may NLP tasks due to LLMs, see e.g. Brown et al. (2020).”
BackgroundInfluentialShojaee, P., Mirzadeh, I., Alizadeh, K., Horton, M., Bengio, S., & Farajtabar, M. (2025) · arXiv · doi:10.48550/arxiv.2506.06941 · Cited by 45
“While these LLMs demonstrate promising language understanding with strong compression capabilities, their intelligence and reasoning abilities remain a critical topic of scientific debate [7, 8].”
EvoPrompting: Language Models for Code-Level Neural Architecture Search
Chen, A., Dohan, D. M., & So, D. R. (2023) · arXiv · doi:10.48550/arxiv.2302.14838 · Cited by 45
Desirable or Distasteful? Exploring Uncertainty in Human-Chatbot Relationships
Pan, S., Cui, J., & Mou, Y. (2024) · International Journal of Human–Computer Interaction · doi:10.1080/10447318.2023.2256554 · Cited by 44
“Significant debate has arisen regarding the application of large language models in designing chatbots, mainly due to worries about the biases they may contain (Bender et al., 2021).”
Woo, D. J., Wang, Y., Susanto, H., & Guo, K. (2023) · Journal of Educational Computing Research · doi:10.1177/07356331231175999 · Cited by 44
“Based on a prompt, an NLG tool generates output text (Bender et al., 2021), which is the tool ’ s prediction of subsequent words, sentences or paragraphs for the prompt (Liu et al., 2021).”
MethodThe Dangers of Underclaiming: Reasons for Caution When Reporting How NLP Systems Fail
Bowman, S. (2022) · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2022.acl-long.516 · Cited by 44
“…it is easy to find references that do not mention the adversarial design of the data and that therefore make claims that are hard to justify: 5 2021) misleadingly describe a form of adversarial data collection 6 as a method for the “careful manipulation of the test data to remove spurious…”
BackgroundMethodInfluentialCoding energy knowledge in constructed responses with explainable NLP models
Gombert, S., Mitri, D. D., Karademir, O., Kubsch, M., Kolbe, H., Tautz, S., Grimm, A., Bohm, I., Neumann, K., & Drachsler, H. (2022) · Journal of Computer Assisted Learning · doi:10.1111/jcal.12767 · Cited by 63
‘Feeding the world, byte by byte’: emergent imaginaries of data productivism
de Wit, M. M., & Canfield, M. C. (2023) · The Journal of Peasant Studies · doi:10.1080/03066150.2023.2232997 · Cited by 56
“…sequences of linguistic forms according to probabilistic information about how they combine, a ‘ stochastic parroting ’ of their vast training data (Bender et al. 2021) These technologies are already showing up in agriculture, and although experts have pointed to their fundamental limitations…”
BackgroundAssessing the Strengths and Weaknesses of Large Language Models
Lappin, S. (2024) · Journal of Logic, Language and Information · doi:10.1007/s10849-023-09409-x · Cited by 45
“Bender et al. (2021) and Chomsky et al. (2023) offer recent versions of this view.”
Can large language models help augment English psycholinguistic datasets?
Trott, S. (2024) · Behavior Research Methods · doi:10.3758/s13428-024-02337-z · Cited by 87
“One key limitation is that most large language models like GPT-4 are trained primarily on written English text produced by a relatively biased sub-set of English speakers (Bender et al., 2021; Chang & Ber-gen, 2023).”
BackgroundGoswami, J., Prajapati, K. K., Saha, A., & Saha, A. K. (2024) · Applied Soft Computing · doi:10.1016/j.asoc.2024.111531 · Cited by 59
Auditing and instructing text-to-image generation models on fairness
Friedrich, F., Brack, M., Struppek, L., Hintersdorf, D., Schramowski, P., Luccioni, A. S., & Kersting, K. (2024) · AI and Ethics · doi:10.1007/s43681-024-00531-5 · Cited by 82
Quo Vadis ChatGPT? From Large Language Models to Large Knowledge Models
Venkatasubramanian, V., & Chakraborty, A. (2024) · Computers and Chemical Engineering · doi:10.1016/j.compchemeng.2024.108895 · Cited by 42
Certification systems for machine learning: Lessons from sustainability
Matus, K. J. M., & Veale, M. (2022) · Regulation & Governance · doi:10.1111/rego.12417 · Cited by 43
“Impacts on human rights through processes of data collection or labeling (Roberts 2019), or impact on emissions from large model training (Bender et al .”
Anticipating Safety Issues in E2E Conversational AI: Framework and Tooling
Dinan, E., Abercrombie, G., Bergman, A. S., Spruit, S., Hovy, D., Boureau, Y.-L., & Rieser, V. (2021) · arXiv · doi:10.48550/arxiv.2107.03451 · Cited by 44
“Notably, the list of words and phrases used to detect which responses contain unsafe language (§5.2.1) contains words like twink ; filtering out or marking these words as “unsafe” may have the effect of limiting discourse in spaces for LGBTQ+ people (Bender et al., 2021).”
BackgroundInfluentialBridging the Gap: A Survey on Integrating (Human) Feedback for Natural Language Generation
Fernandes, P., Madaan, A., Liu, E., Farinhas, A., Martins, P. H., Bertsch, A., de Souza, J. G. C., Zhou, S., Wu, T., Neubig, G., et al. (2023) · arXiv · doi:10.48550/arxiv.2305.00955 · Cited by 44
Resistance and refusal to algorithmic harms: Varieties of ‘knowledge projects’
Ganesh, M. I., & Moss, E. (2022) · Media International Australia · doi:10.1177/1329878x221076288 · Cited by 70
“A different set of dynamics emerges through the scandal at Google ’ s Ethical AI Lab that resulted in the termination of the Lab ’ s co-founders; this began over a paper that outlined how Google and its engineers constitute a site of ethical decision-making (Bender et al., 2021).”
BackgroundDetecting Intra-Field Variation in Rice Yield With Unmanned Aerial Vehicle Imagery and Deep Learning
Bellis, E. S., Hashem, A. A., Causey, J., Runkle, B. R. K., Moreno‐García, B., Burns, B. W., Green, V. S., Burcham, T. N., Reba, M. L., & Huang, X. (2022) · Frontiers in Plant Science · doi:10.3389/fpls.2022.716506 · Cited by 51
“Compared to 3D-CNNs, 2D-CNNs would also require less computational and environmental resources for training (Strubell et al., 2019; Henderson et al., 2020; Bender et al., 2021) and fewer flyovers to generate yield predictions when models are deployed.”
BackgroundBeware of sustainable AI! Uses and abuses of a worthy goal
Heilinger, J., Kempt, H., & Nagel, S. K. (2023) · AI and Ethics · doi:10.1007/s43681-023-00259-8 · Cited by 69
“Bender et al. point towards this tension in their paper on “stochastic parrots” [3] in a striking way: not only do they question the social sustainability of large language models by pointing towards issues of how these language models came about, but they also question the ends at which their…”
Who Determines What Is Relevant? Humans or AI? Why Not Both?
Faggioli, G., Dietz, L., Clarke, C. L. A., Demartini, G., Hagen, M., Hauff, C., Kando, N., Kanoulas, E., Potthast, M., Stein, B., et al. (2024) · Communications of the ACM · doi:10.1145/3624730 · Cited by 53
Adaptable and Interpretable Neural MemoryOver Symbolic Knowledge
Verga, P., Sun, H., Baldini Soares, L., & Cohen, W. (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.288 · Cited by 41
“all of the underlying biases within that data (Zhao et al., 2019; Bender et al., 2021).”
BackgroundIntegrating Data Science Ethics Into an Undergraduate Major: A Case Study
Baumer, B. S., Garcia, R. L., Kim, A. Y., Kinnaird, K. M., & Ott, M. Q. (2022) · Journal of Statistics and Data Science Education · doi:10.1080/26939169.2022.2038041 · Cited by 62
LLMs for Financial Advisement: A Fairness and Efficacy Study in Personal Decision Making
Lakkaraju, K., Jones, S. E., Vuruma, S. K. R., Pallagani, V., Muppasani, B. C., & Srivastava, B. (2023) · 4th ACM International Conference on AI in Finance · doi:10.1145/3604237.3626867 · Cited by 42
“In [3], the authors argue that while these models have pushed the boundaries of what is possible through architectural innovations and sheer size, there are potential risks associated with their use.”
BackgroundGitTables: A Large-Scale Corpus of Relational Tables
Hulsebos, M., Demiralp, Ç., & Groth, P. (2023) · Proceedings of the ACM on Management of Data · doi:10.1145/3588710 · Cited by 129
“Besides information spread it is known that training models on large uncurated datasets may impose biases [3].”
BackgroundDeceptive AI Ecosystems: The Case of ChatGPT
Zhan, X., Xu, Y., & Sarkadi, S. (2023) · Proceedings of the 5th International Conference on Conversational User Interfaces · doi:10.1145/3571884.3603754 · Cited by 43
“Like the risks associated with ‘stochastic parrots’ that have been proposed by Bender et al. [13], while LLMs are remarkable for their ability to generate text that emulates human composition, the fact that these texts lack genuine meaning behind is a very disturbing thought.”
BackgroundBridging the Gap: A Survey on Integrating (Human) Feedback for Natural Language Generation
Fernandes, P., Madaan, A., Liu, E., Farinhas, A., Martins, P. H., Bertsch, A., de Souza, J. G. C., Zhou, S., Wu, T., Neubig, G., et al. (2023) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00626 · Cited by 44
Spennemann, D. H. R. (2024) · Heritage · doi:10.3390/heritage7030070 · Cited by 58
Is AI my co-author? The ethics of using artificial intelligence in scientific publishing
Moffatt, B., & Hall, A. (2024) · Accountability in Research · doi:10.1080/08989621.2024.2386285 · Cited by 60
“In many ways, these AI models are better thought of as confabulation engines mimicking common human speech (Bender et al. 2021).”
How to train your stochastic parrot: large language models for political texts
Ornstein, J. T., Blasingame, E. N., & Truscott, J. S. (2025) · Political Science Research and Methods · doi:10.1017/psrm.2024.64 · Cited by 44
“2 We conclude with a discussion on whether the performance gains we document are worth the potential dangers associated with unprincipled use of these models (Strubell et al. , 2019; Abid et al. , 2021; Bender et al. , 2021; Spirling, 2023).”
InfluentialPeykani, P., Ramezanlou, F., Tănăsescu, C. R., & Ghanidel, S. (2025) · Applied Sciences · doi:10.3390/app15148103 · Cited by 85
Generative AI for Requirements Engineering: A Systematic Literature Review
Cheng, H., Husen, J. H., Lu, Y., Racharak, T., Yoshioka, N., Ubayashi, N., & Washizaki, H. (2025) · Software Practice and Experience · doi:10.1002/spe.70029 · Cited by 171
“Early industrial experiments with RAG and hybrid approaches have shown that grounding LLM outputs in domain-specific knowledge can markedly improve reliability, reduce hallucinations, and even support multilingual requirement processing.”
InfluentialSubspace Regularizers for Few-Shot Class Incremental Learning
Akyürek, A. F., Akyürek, E., Wijaya, D. T., & Andreas, J. (2021) · arXiv · doi:10.48550/arxiv.2110.07059
“…to recognize new concepts (e.g. new skills in personal assistants or new road signs in self-driving vehicles), while training data is sometimes unavailable for reuse (e.g. due to privacy regulations, Lesort et al. 2019; McClure et al. 2018; or storage and retraining costs, Bender et al. 2021).”
BackgroundLund, B. D., Wang, T., Mannuru, N. R., et al. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4389887 · Cited by 42
On the Evolution of Knowledge Graphs: A Survey and Perspective
Jiang, X., Xu, C., Shen, Y., Sun, X., Tang, L., Wang, S., Chen, Z., Wang, Y., & Guo, J. (2023) · arXiv · doi:10.48550/arxiv.2310.04835 · Cited by 42
Galke, L. P. A., & Scherp, A. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-long.279 · Cited by 78
Whose Language Counts as High Quality? Measuring Language Ideologies in Text Data Selection
Gururangan, S., Card, D., Dreier, S., Gade, E., Wang, L., Wang, Z., Zettlemoyer, L., & Smith, N. A. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.165 · Cited by 42
“The ethical and legal norms around scraping public-facing web data, especially those produced by minors, are still in flux (Fiesler et al., 2020), and may not align with user perceptions of what constitutes fair use of online communications (Williams et al., 2017).”
BackgroundWhy large language models are poor theories of human linguistic cognition: A reply to Piantadosi
Katzir, R. (2023) · Biolinguistics · doi:10.5964/bioling.13153 · Cited by 89
“But current LLMs remain the stochastic parrots that Bender et al. (2021) tell us they are.”
BackgroundGenerative Artificial Intelligence: Fundamentals
Corchado, J. M., Sebastian López, F., V., J., Raul Garcia, S., & Chamoso, P. (2023) · ADCAIJ ADVANCES IN DISTRIBUTED COMPUTING AND ARTIFICIAL INTELLIGENCE JOURNAL · doi:10.14201/adcaij.31704 · Cited by 62
Managing the Twin Faces of AI: A Commentary on “Is AI Changing the World for Better or Worse?”
Shankar, V. (2024) · Journal of Macromarketing · doi:10.1177/02761467241286483 · Cited by 41
“Gen AI models or LLMs only probabilistically link words and sentences without considering their meaning and can make up answers, prompting researchers to term them “ stochastic parrots ” (Bender et al. 2021).”
Applications of Large Language Models in the Field of Suicide Prevention: Scoping Review
Holmes, G., Tang, B., Gupta, S., Venkatesh, S., Christensen, H. M., & Whitton, A. E. (2025) · Journal of Medical Internet Research · doi:10.2196/63126 · Cited by 80
Zheng, G., & Brintrup, A. M. (2025) · International Journal of Production Research · doi:10.1080/00207543.2025.2543964 · Cited by 82
Generative AI and English language teaching: A global Englishes perspective
Lee, S., Jeon, J., McKinley, J., & Rose, H. (2025) · Annual Review of Applied Linguistics · doi:10.1017/s0267190525100184 · Cited by 104
“Furthermore, LLMs’ statistical algorithms often result in homogenization simplifying linguistic diversity and excluding less-resourced languages and their associated cultures (Bender et al., 2021; Dai et al., 2024).”
PERFICT: A Re‐imagined foundation for predictive ecology
McIntire, E. J. B., Chubaty, A. M., Cumming, S. G., Andison, D., Barros, C., Boisvenue, C., Haché, S., Luo, Y., Micheletti, T., & Stewart, F. E. C. (2022) · Ecology Letters · doi:10.1111/ele.13994 · Cited by 50
“The diminishing returns that may come from iterative improvements for any specific model should, however, be compared against costs of, e.g., particularly large models (Bender et al., 2021).”
BackgroundA taxonomy and review of generalization research in NLPPreprint
Hupkes, D., Giulianelli, M., Dankers, V., Artetxe, M., Elazar, Y., Pimentel, T., Christodoulopoulos, C., Lasri, K., Saphra, N., Sinclair, A., et al. (2022) · arXiv (Cornell University) · doi:10.3929/ethz-b-000638992 · Cited by 41
Augmenting research methods with foundation models and generative AI
Rossi, S., Rossi, M., Mukkamala, R. R., Thatcher, J. B., & Dwivedi, Y. K. (2024) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2023.102749 · Cited by 66
Collecting a large-scale gender bias dataset for coreference resolution and machine translation
Levy, S., Lazar, K., & Stanovsky, G. (2021) · Findings of the Association for Computational Linguistics: EMNLP 2021 · doi:10.18653/v1/2021.findings-emnlp.211 · Cited by 41
MOMENT: a family of open time-series foundation models
Goswami, M., Szafer, K., Choudhry, A., Cai, Y., Li, S., & Dubrawski, A. (2024) · arXiv · doi:10.48550/arxiv.2402.03885 · Cited by 41
Leveraging generative AI for clinical evidence synthesis needs to ensure trustworthiness
Zhang, G., Jin, Q., Jered McInerney, D., Chen, Y., Wang, F., Cole, C. L., Yang, Q., Wang, Y., Malin, B. A., Peleg, M., et al. (2024) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2024.104640 · Cited by 40
Ehsan, U., Singh, R., Metcalf, J., & Riedl, M. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533186 · Cited by 40
FaithDial: A Faithful Benchmark for Information-Seeking Dialogue
Dziri, N., Kamalloo, E., Milton, S., Zaiane, O., Yu, M., Ponti, E. M., & Reddy, S. (2022) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00529 · Cited by 63
“By mimicking the distributional properties of the data, models are bound to ‘‘parrot’’ the hallucinated signals at test time (Bender et al., 2021).”
BackgroundBano, M., Hoda, R., Zowghi, D., & Treude, C. (2023) · Automated Software Engineering · doi:10.1007/s10515-023-00407-8 · Cited by 84
“Despite preliminary progress, there remains a significant lack of clarity on how LLMs compare to human intelligence in qualitative research (Bender et al. 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.