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 5,801–5,900 of 6,931
Illusionism About Phenomenal Consciousness
Kammerer, F. (2026) · Oxford University Press eBooks · doi:10.1093/9780197840283.003.0003
Shedding Our Cartesian Scales: Illusionism and Epistemology
Kammerer, F. (2026) · Oxford University Press eBooks · doi:10.1093/9780197840283.003.0011
Meet the AI Moment: Using Afrofuturism and Technowomanism Together as Technological Practice
Bosley, B., Klassen, S., Harrington, C., Asare, I., Almeida, F., Idaewor-Bless, H., C, M., & Salih, Y. (2026) · Designing Interactive Systems Conference · doi:10.1145/3802974.3808007
34115 Machine learning in communication science: from word frequencies to large language models
Viehmann, C. (2026) · De Gruyter eBooks · doi:10.1515/9783111203577-015 · Cited by 1
Yang, Y., Meng, F., Nah, F. F., & Lc, R. (2026) · Designing Interactive Systems Conference · doi:10.1145/3800645.3812950 · Cited by 1
EduReview: An AI–assisted Risk Aware Design Critique System for K–12 Products
Liang, Q. (2026) · Designing Interactive Systems Conference · doi:10.1145/3802974.3809477
Queer Zineographies: Materializing Tactics for Resisting AI and Data Systems
Riggs, A. T., Meyer, L. S., O'Reilly-Kime, M., Armstrong, T., Kender, K., Osipova, E., Tran, A.-T., Taylor, J., Rothschild, A., Grabe, I., et al. (2026) · Designing Interactive Systems Conference · doi:10.1145/3800645.3812834 · Cited by 1
“…data systems focus on both critiquing and deconstructing the technology’s harms that exacerbate the vulnerability of already marginalized groups [17, 44, 62, 84, 103], as well as exploring queer engagements with AI, such as through artistic or autoethnographic reflections [11, 78, 83, 141, 142].”
Implicit Affective Steering: A Design Probe for Generative Curatorial Storytelling
Ma, M. C., Shi, L., Li, H., Cheng, J., & Kirk, D. S. (2026) · Designing Interactive Systems Conference · doi:10.1145/3800645.3812861
32715 Content analysis in quantitative communication research
Andreeva, I. N., Taggart, M., Lattimer, T. A., & Hahn, L. (2026) · De Gruyter eBooks · doi:10.1515/9783111087139-015 · Cited by 1
Stepping Into the Black Box: Opening Up LLMs to Public Exploration Through Discursive Design
Immel, S., Rajani, N., Verweij, R., Li, J., Nissen, B., Soares, L. L., & Taylor, A. S. (2026) · Designing Interactive Systems Conference · doi:10.1145/3800645.3812976 · Cited by 1
“However, we left bits of exposed wooden structure and decorated the inside with wire, pieces of broken machinery, and repurposed 3D printer waste to convey some sense of the imperfection and materiality of LLM systems [4].”
Ethical Foundations of AI in Social Work
An, R., & Lindsey, M. A. (2026) · Springer eBooks · doi:10.1007/978-3-032-18443-6_2
The Case for Repeatable, Open, and Expert-Grounded Hallucination Benchmarks in Large Language Models
Norman, J., Rivera, M. U., & Hughes, D. A. (2026) · International Conference on Information and Computer Technologies (ICICT) · doi:10.1145/3803291.3803328 · Cited by 1
“Incoherence Current generation language models produce syntactically correct, seemingly coherent strings of text [4].”
InfluentialA conversational AI toolkit for citizen deliberation in party primary elections
Jiménez, F. R. V., & Malý, M. (2026) · Open Research Europe · doi:10.12688/openreseurope.24102.1
AI-designed versus pre-service teacher-designed activities: Analysis and evaluation
Kır, D. B. (2026) · Innovations in Education and Teaching International · doi:10.1080/14703297.2026.2689673
“This partly aligns with Bender et al.’s (2021) critique of large language models as ‘stochastic parrots’, which cautions that models recombine existing patterns rather than generate genuinely novel products.”
Framing the Future: The Impact of AI-Generated Video on Major Film Creators
White, H. (2026) · Unveiling seven continents yearbook journal · doi:10.65326/u7y566816
Sokolov, A. V. (2026) · Issues of Media Business · doi:10.65324/imb040
Demin, P. E. (2026) · doi:10.37539/260430.2026.53.78.013
SAFE AI: Designing a Gender-Responsive and Child-Safe AI Framework
Rosalinda, H., Liliana, D. Y., & Pallikadavath, S. (2026) · doi:10.1145/3816713.3819509
Beyond Over-Editing: Important Weight Constrained Knowledge Editing in Large Language Models
Zhu, Z. G., Dai, Y., Guo, B., Zheng, H., Sun, Y., & Fan, Z. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810656
Zachariah, A., Simkin, B., Pope, N. D., et al. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810435
Design Methodology and Performance Trade-offs Management for Distributed and Compound AI Systems
Gravara, M., Stanisic, A., & Nastic, S. (2026) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud72782.2026.00049 · Cited by 3
“Improving performance for a given task relies primarily on scaling the model through additional parameters, training data, or compute budget, following empirical scaling laws [4], [8], [9].”
Laabichi, S., & Bahmad, M. (2026) · SHS Web of Conferences · doi:10.1051/shsconf/202623207032
Thinking Outside the Black Box: Seven Guidelines for GenAI Use in the Ecological Sciences
Ferrari, N. C., Conklin, E. E., Fitch, A. A., Gannon, D. G., Macedo, R., Place, A., Sachs, H. R., Seo, E., Weldy, M. J., & Betts, M. G. (2026) · Ecology and Evolution · doi:10.1002/ece3.74118
“For instance, the output of LLMs is better understood as a recombination of patterns from training data (i.e., “stochastic parroting”) than as the generation of factual information to learn from (Bender et al. 2021; Ji et al. 2023; Kalai et al. 2025).”
Risk-Aware Medical Question Answering as a Reliability-Critical Decision Process
Luo, Y., & Li, Y. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810784
The new era in financial evolution: a review of applications of generative AI in finance
Feng, S., Yeo, E. X. Y., Ong, R., Ren, J. S. J., Tan, Y. L., Xie, M., Wang, J., & Li, F. (2026) · Memetic Computing · doi:10.1007/s12293-026-00512-7
Mart, M., & Kaya, G. (2026) · International Journal of Early Childhood · doi:10.1007/s13158-026-00526-9
Brian, M. R. (2026) · Educational Research and Reviews · doi:10.5897/err2026.4511
Designing Safe Hybrid AI–XR Simulations for Healthcare Communication and Interaction Training
Mikkonen, K., Butkevičiūtė, E., Armalis, P., et al. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-28812-7_21
Wirth, U. (2026) · Zeitschrift für Weiterbildungsforschung · doi:10.1007/s40955-026-00350-7
Ma, X., Zhu, R., Wang, Z., Xiong, J., Chen, Q., Camp, L. J., & Tang, H. (2026) · Journal of Medical Internet Research · doi:10.2196/89003
Using AI-assisted coding to build digital tools for natural history collections
de Brito, V. (2026) · Natural History Collections and Museomics · doi:10.3897/nhcm.3.191202
“Large language models are stochastic by design, gen - erating words (code, in this case) by sampling from a probability distribution (Bender et al. 2021).”
Linh, V. T. (2026) · Proceedings of International Academia · doi:10.65956/procia.2026.113
“The student ’ s remark about sometimes blindly accepting AI suggestions points to the risk of over-reliance and superficial processing (Bender et al., 2021).”
Zhang, W., Dizon, D., & Shin, Y. (2026) · Research Square · doi:10.21203/rs.3.rs-9805028/v1
ESPOSITO, N., Tricarico, A., Porzio, L., Ardebili, A. A., & Stella, M. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6955453
Who is the muse? Prompt writing, AI poets, and the sympoietic author-function
Chauhan, V. (2026) · AI & Society · doi:10.1007/s00146-026-03147-x · Cited by 2
“Models such as GPT and Gemini are trained on datasets that include literature, journalism, online forums, and academic texts, enabling them to internalize syntactic, semantic and stylistic patterns (Bender et al. 2021; Audichya and Saini 2023; Xu et al. 2025).”
Seidu, A. R. A. (2026) · Discover Energy · doi:10.1007/s43937-026-00165-0
Albataineh, J. (2026) · SN Social Sciences · doi:10.1007/s43545-026-01531-4
“Analyses of GPT-family systems note their capacity for fluent multi-turn explanation, but also their prompt sensitivity and potential to generate plausible but incorrect output (Floridi and Chiriatti 2020; Bender et al. 2021).”
Patnaik, T., & Nanda, N. S. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7036-1.ch003
Rita, G. A. F., Melo, L. G. M. D. C., & Moreira, M. L. D. L. (2026) · RECIMA21 - Revista Científica Multidisciplinar · doi:10.47820/recima21.v7i6.8322 · Cited by 2
Chen, X., & Fadzil, K. S. (2026) · Global Philosophy · doi:10.1007/s10516-026-09790-9
“The second position is epistemic dismissalism: the view that AI outputs are epistemically worthless objects, the product of complex pattern matching, and not belonging to any epistemically relevant category (Bender et al. 2021).”
Gupta, S., Bhatia, K., Gulkotwar, N., Sharma, R. R., & Kaur, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6960803
I Want an Assistant: Generative AI, Women’s Leadership, and the Redistribution of Invisible Labor
Kostihová, M., & Makarevitch, I. (2026) · Merits · doi:10.3390/merits6020017
“…theorized in various parallel ways, as the “stochastic parrot” which amplifies hegemonic biases engrained in training data with apparent coherence [55], or as the “Smart Wife”, a feminized, anticipatory, and endlessly accommodating digital assistant embedded in the design and use of smart…”
Kalām, Humans and AI: Reason(ing), Creation/Creativity, and Agency
Guessoum, N. (2026) · Religions · doi:10.3390/rel17060703 · Cited by 1
Recent Advances and Challenges in Natural Language Processing
Kalra, A., & Patil, L. N. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-3361-8.ch001
Generative AI adoption and agile management systems: insights from HR managers
Dey, C. (2026) · Management Research The Journal of the Iberoamerican Academy of Management · doi:10.1108/mrjiam-01-2026-1905
Mutual supervision: A framework for improving reward model in preference alignment
Peng, Z., Zhang, M., Wen, G., Zhang, Q., & Chen, C. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.116467 · Cited by 1
Potential and limitations of LLMs for analyzing narratives on corporate profits
Bonhoure, É., Bordea, G., Doucet, A., Fabre, A., Hamdi, A. T. E., Hendieh, J., Labardin, P., Moreno, J. G., & Riva, A. (2026) · Revue de l’organisation responsable · doi:10.54695/ror.212.0098
Colborn-Clark, D., Cecchinato, M. E., & Dow, A. (2026) · Annual Symposium on Human-Computer Interaction for Work (CHIWORK) · doi:10.1145/3808045.3808057
Rothschild, A., Nematova, M. K., Eickman, B., DiSalvo, C. F., & DiSalvo, B. J. (2026) · Annual Symposium on Human-Computer Interaction for Work (CHIWORK) · doi:10.1145/3808045.3808052
XAI2Brain: A Perspective on Mechanistic Interpretability for Brain–AI Alignment
Jiang, R., Zhou, Y., Wang, B., Angelov, P. P., & Ni, Q. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8060167
“As LLMs display increasingly complex behaviors, the line between programmed responses and genuine cognitive processes becomes blurred, presenting a unique challenge for XAI in demarcating clear explanations for AI actions that mimic conscious decisions [64,90].”
Addressing the Challenges in Using Synthetic Data for Health Research: Application to Cardiology
Baschet, L., Marque, S., Locret, S., Jensen, J., Barbet, V., & Patrick Jourdain, M. (2026) · JMIR Cardio · doi:10.2196/92930
From Realism to Relevance: Rethinking Simulator Fidelity in the Era of Generative AI
Schoenherr, J. R. (2026) · IEEE Transactions on Technology and Society · doi:10.1109/tts.2026.3693463
Syharat, C. M., Zaghi, A. E., & Motaref, S. (2026) · Frontiers in Research Metrics and Analytics · doi:10.3389/frma.2026.1830196
“These concerns draw on analyses of environmental and social impacts associated with large-scale AI systems (Bender et al., 2021; Luccioni et al., 2023), even as some scholars note the potential for AI applications to contribute to energy optimization and climate mitigation (Patterson et al., 2021;…”
Unseen ethical paradoxes of inclusive education in the age of artificial intelligence
Kohout‐Diaz, M. (2026) · International Journal of Inclusive Education · doi:10.1080/13603116.2026.2691584 · Cited by 1
ATLAS: Harnessing retrieval-augmented generation (RAG)
Smithies, J., Berman, G., Taiuru, K., Spychal, M., & Moore, J. (2026) · Cambridge Forum on AI Culture and Society · doi:10.1017/cfc.2026.10029
“…parrots” article “… an LM [language model] is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning …” (Bender et al. 2021, 617).”
Kim, R. S. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01573-w
“This complements known RLHF failure modes (e.g., reward misgeneralization, superficial politeness) [15, 16] and adds a distinct, affective brittleness.”
Sun, H., Cheng, W., Pae, H. K., & Wang, L. (2026) · International Journal of Applied Linguistics · doi:10.1111/ijal.70258 · Cited by 1
Consciousness, creativity, and understanding are not obstacles to machine intelligence
Pesonen, R., & Reijula, S. (2026) · Synthese · doi:10.1007/s11229-026-05655-1
“This time, the metaphor is not the Chinese room, but a stochastic parrot (Bender et al., 2021).”
Ayan, M. A. (2026) · Gazi Üniversitesi Fen Bilimleri Dergisi Part C Tasarım ve Teknoloji · doi:10.29109/gujsc.1924721
Sormani, P., Ikegami, T., Saunier, A., Jobin, A., Glassey, O., & Salter, C. L. (2026) · Cambridge Forum on AI Culture and Society · doi:10.1017/cfc.2026.10024
Bamil, V. (2026) · IGI Global eBooks · doi:10.4018/979-8-3373-1509-6.ch003
Synthetic resonance: a framework for growth-oriented human–AI relationships
Fabes, R. A. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1891738
“AI systems generate responses through pattern-based modeling, producing outputs that align with user expectations without relying on subjective experience (Bender et al., 2021).”
FLEX your MBA: Empowering learning efficiency in the digital classroom using an AI‐based application
Shah, P., Mishra, I., & Amarasuriya, S. (2026) · Decision Sciences Journal of Innovative Education · doi:10.1111/dsji.70024
Puutio, T. A., & Do, M. (2026) · Group Decision and Negotiation · doi:10.1007/s10726-026-10006-z · Cited by 1
“A growing body of research is consistent with LLMs being capable of reproducing not only surface-level task per - formance, but also deeper psychological regularities, including heuristic reasoning, framing effects, and systematic deviations from normative rationality (Bender et al., 2021).”
On Technology and the Ecology of Thinking: Thinking and Aliveness in the Brave New World
Selek, M. (2026) · International Journal of Applied Psychoanalytic Studies · doi:10.1002/aps.70060 · Cited by 2
Andrade, J. G. V., Andrade, M. Á. V., Zurita, D. E. G., & Borja, X. F. T. (2026) · Revista Científica Multidisciplinar G-nerando · doi:10.66473/rcmg.v7i1.1238
Penda, I. A. (2026) · Journal Of Organizational Behavior Research · doi:10.51847/s5toitp22m
Chintacunta, S. N. R., & Deva, S. (2026) · European Journal of Artificial Intelligence and Machine Learning · doi:10.24018/ejai.2026.5.3.70149
Positioning value: mapping the logics of persuasion in Science, with evidence from AI research
Tamang, T. (2026) · Theory and Society · doi:10.1007/s11186-026-09710-1
“…value is elevated for its theoretical importance, recognizing it as a modern application of Occam’s Razor, a long-standing principle favoring simplicity in model selection to improve generalization and interpretability (Blumer et al., 1987; Domingos, 1999 (Bender et al., 2021; Gebru et al., 2021).”
Algorithmic Bias in AI Systems: Examining Its Impact on Fairness and Decision-MakingPreprint
Patel, D. (2026) · Research Square · doi:10.21203/rs.3.rs-10083731/v1
Wölfle, A., & Schmuck, D. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2685242
Kelly, J. T. (2026) · Springer eBooks · doi:10.1007/978-3-032-26782-5_3
Developing emotional intelligence through affective outdoor learning
Boyes, M., McNatty, S., & Parker, S. (2026) · Journal of Adventure Education & Outdoor Learning · doi:10.1080/14729679.2026.2686120 · Cited by 1
“Presently, they lack accuracy and reliability (Bender et al., 2021) and can make wrong predictions (Deckker & Sumanasekara, 2025).”
The LLM Mirage: Economic Interests and the Subversion of Weaponization Controls
Gupta, R., & Reddie, A. W. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806538 · Cited by 1
Gautam, S., Liu, H., Choi, Y., & Lease, M. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812262 · Cited by 2
“Our study was reviewed and approved by our organization’s Institutional Review Board (IRB).”
Sande, N. (2026) · Verbum et Ecclesia · doi:10.4102/ve.v47i1.3720
Deva, R., Thapa, A., Mehta, Z., Kapile, S. K., Jalota, S., Karusala, N., & Ismail, A. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806721 · Cited by 1
“Given the well-documented risks associated with LLMs—such as misinformation, hallucinations, and bias [9, 23, 36, 54]—it is critical to evaluate such applications rigorously before deployment.”
Vonschallen, S., Thür, A., Schmiedel, T., & Eyssel, F. (2026) · Computers in Human Behavior Artificial Humans · doi:10.1016/j.chbah.2026.100337
The meek shall inherit the Earth: personality, empathy, and the human shaping of LLMs
Mangalam, M. (2026) · AI and Ethics · doi:10.1007/s43681-026-01215-y
“We do not claim that AI systems possess internal experiences, beliefs, or developmental trajectories equivalent to human cognition; LLMs remain statistical models that predict token sequences based on learned patterns [12].”
Malloy, T., Wax, J., Bissyandé, T. F., & Klein, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6985745
Co-Constructing Alignment: A Participatory Approach to Situating AI Values
Arzberger, A., Liscio, E., Lupetti, M. L., de Troya, Í. M. D. R., & Yang, J. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812267 · Cited by 4
Neomythism as Postdigital Pedagogy
de Lune, R. (2026) · Postdigital Science and Education · doi:10.1007/s42438-026-00677-x
Hudig, A. I., & Singh, J. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806741
Reckoning with the Political Economy of AI: Avoiding Decoys in Pursuit of Accountability
Vertesi, J., boyd, D., Taylor, A. S., & Shestakofsky, B. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806739 · Cited by 7
AlDahoul, N., Karim, H. B. A., & Tan, M. J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6900538
AI-ClassSims: Multi-Agent LLM Classroom Simulation with Real-Time Feedback for Teacher Training
Memari, M., & Ruggles, K. (2026) · Intermountain Engineering, Technology and Computing (IETC) · doi:10.1109/ietc69527.2026.11568604
Artificial General Intelligence and Public Diplomacy: A Scenario Analysis of Possible FuturesPreprint
Athar, M. M. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6985150
Шпак, О., & Сиротич, О. (2026) · Physico-mathematical modelling and informational technologies · doi:10.15407/fmmit2026.42.109
Jordan, D., Fonseka, D., & Sitvaev, J. S. A. D. (2026) · IntechOpen eBooks · doi:10.5772/intechopen.1016141
Mahajan, P. T. (2026) · AI and Ethics · doi:10.1007/s43681-026-01202-3 · Cited by 1
Legitimate: Resurrecting Biased Ontologies On the Path to Superintelligence
Monroe‐White, T., White, K., & Hashmi, N. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806757
“It fails at multi-step reasoning and mathematics [52], produces confident yet incorrect responses [49, 82], and routinely reproduces embedded biases from its training corpora and the model optimization choices of its human creators [6, 48].”
Don't Trust the Process: When Verifiability Undermines AI Accountability
Paris, T., Moon, A., & Guo, J. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812373 · Cited by 1
Greek Mythology as a Lens for Understanding the Risks, Biases, and Ethics of Artificial Intelligence
Muhlenbach, F. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812367
The Stability Trap: Evaluating the Reliability of LLM-Based Instruction Adherence Auditing
Shergadwala, M. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806536 · Cited by 1
“…often masks underlying reasoning failures, particularly in quantitative syntactic tasks; and (3) We discuss a Hybrid Auditing Protocol for enterprise governance, recommending that all deterministically verifiable logic be delegated to code, while LLMs are reserved for complex semantic evaluation.”
Li, Y., & Taneja, H. (2026) · Communication and Change · doi:10.1007/s44382-026-00030-7 · Cited by 2
“At the level of model development, pretraining on large text corpora and subsequent alignment and safety fine-tuning constrain which topics and framings are more readily produced (Abid et al., 2021; Bender et al., 2021).”
A Gender-Diverse Perspective of Bias in Large Language Models
Gaba, A., Wall, E., Keyes, O., Brun, Y., Hall, K. W., & Xiong, C. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812257
“As these systems become more integrated into daily life, concerns about bias have intensified [7, 60].”
Rehak, R., & Woodcock, T. K. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812357
“First, the majority of narratives regarding AI systems largely overstate their capabilities [7, 52], and particularly in current conflicts, it is challenging to access valid empirical data, such as actual system configurations or technical manuals, due to both corporate and military secrecy.”
Lyu, C., & Wu, F. (2026) · Journal of Broadcasting & Electronic Media · doi:10.1080/08838151.2026.2684553
“This increased variability may partly reflect the stochastic nature of generative AI systems (Bender et al., 2021), whereas human-written content is shaped by editorial norms that promote consistency and moderation.”
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.