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 6,601–6,700 of 6,931
Epistemic Disobedience by Design: Leveraging GPT to Pluralize Architectural Knowledge Systems
Roslan, S. B., Ramli, H., & Yi, K. F. (2026) · Springer eBooks · doi:10.1007/978-981-92-3527-8_12
Wilson, J. (2026) · Journal of Linguistic Anthropology · doi:10.1111/jola.70074
Framing AI as a Learning Tool and Reassessing Learning
Popa, E. (2026) · Human Affairs · doi:10.1515/humaff-2026-0078
“Another important critique, particularly regarding generative AI, is their description as “ stochastic parrots ” by Bender et al. (2021).”
Woodworth, J. (2026) · Digital studies in language and literature · doi:10.1515/dsll-2026-0035
“…systems tend to privilege dominant varieties, genres, and rhetorical patterns derived from training data; when their outputs are treated as neutral standards, multilingual features can be marked as de fi cient or in need of normalization (Bender et al. 2021; Canagarajah 2012; García and Li 2014).”
InfluentialBecoming the “post-reality”: AIGC, memory generation, and an experimental foray into AI anthropology
Wu, E., & Ni, M. (2026) · Big Data & Society · doi:10.1177/20539517261481725
219Chapter 11 Means of Discrimination. A Feminist Materialist Analysis of Making and Collecting Data
Savić, S. (2026) · De Gruyter eBooks · doi:10.1515/9783111373096-012
Imagining and Enacting Futures: The Anticipatory Politics of Algorithmic Governance
Chan, N. K., Su, C. C., & Zhou, R. (2026) · Sociology Compass · doi:10.1111/soc4.70249
Zhaı, X., Bewersdorff, A., Lee, G.-G., Feldman-Maggor, Y., Latif, E., & Tai, R. H. (2026) · International Journal of Science Education · doi:10.1080/09500693.2026.2727692 · Cited by 1
“…we proposed a four-layer architecture that embeds human responsibility across the AI lifecycle: The Data Infrastructure layer treats data not as neutral inputs but as socially constructed representations that must be curated for epistemic integrity, inclusivity, and privacy (Bender et al., 2021).”
Mythemes and mythopoesis for transformers and transformation
Macher, B. (2026) · Human Organization · doi:10.1080/00187259.2026.2716813
ScholarGuard-AI An Intelligent Framework for Research Paper Generation and Authenticity Detection
Karthik, S., Kumar, T., Kumari, S., & Kumar, P. (2026) · ITM Web of Conferences · doi:10.1051/itmconf/20269002013
“Generalisation across domains was a challenge for detection and generation systems [17].”
AI-mediated cross-cultural translation of Chinese brush-and-ink psychology for aesthetic empathy
Bei, T., Yan, H., Li, Y., et al. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1828438
“Bender et al.’s (2021) critique of AI art ethics and translational processes has galvanized a systemic rethinking of technological and social justice.”
Chiodo, S. (2026) · AI and Ethics · doi:10.1007/s43681-026-01346-2
LLM-Based Chatbots and the Suicidal Crisis: A Bi-Logical Account of Relational Hallucination
Ruggieri, R. A., Ragosta, A., & Balcombe, L. (2026) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph23091186 · Cited by 2
Disambiguating Adaptivity in AI Research and RegulationPreprint
Radomski, B. M. (2026) · doi:10.14293/pr2199.004442.v1
Word Counting Is Not Enough: A Syntax-Driven Computational Method for Analyzing Participation
Raveau, M. P., Goñi, J., Amigo, P., & Fuentes-Bravo, C. (2026) · Computation · doi:10.3390/computation14090209
Assembling Topic Models: Material Political Economy and the Genealogy of an Algorithm
Zhang, B. (2026) · Social Studies of Science · doi:10.1177/03063127261469570
The BS meter: detecting politics and labour through ChatGPT’s language
Trevisan, A., Giddens, H., Dillon, S., & Blackwell, A. F. (2026) · AI & Society · doi:10.1007/s00146-026-03238-9
“…linguistic critique of computational language models reported by Emily Bender, Timnit Ge-bru and colleagues in their Stochastic Parrots critique (Bender et al., 2021) is a paper by Sp ¨ arck Jones (Sp ¨ arck Jones, 2004), raising questions similar to our own, and also in the spirit of…”
Papu-Zamxaka, V., & Khosa, D. (2026) · Genealogy · doi:10.3390/genealogy10040132
Patault, E., Sinsin, T., Bruis-Gervasone, L., Quesnot, P., Girard, S., Pesquier, B., & MARDUEL, B. (2026) · AI for Engineering · doi:10.3390/aieng1020011
Detecting the undetectable? Reassessing academic misconduct procedures in the era of generative AI
Teichmann, F. M. J. (2026) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2026.2728036
Difficulty as method: A framework for humanities pedagogy in the age of generative AI
Gusain, R. (2026) · Arts and Humanities in Higher Education · doi:10.1177/14740222261487617
Yoo, I. H., & Kang, M. (2026) · The Journal of Korean Association of Computer Education · doi:10.32431/kace.2026.29.8.010
Mnguni, L. E., Nuangchalerm, P., Sari, I. J., Ramulumo, M., Camara, J. S., Biên, N. V., & Islami, R. A. Z. E. (2026) · Discover Education · doi:10.1007/s44217-026-02140-2
“These environmental burdens and the location of such facilities frequently follow established patterns of environmental injustice, in which historically marginalised communities bear disproportionate ecological costs [8–10].”
Shriber, M. (2026) · Poetics Today · doi:10.1215/03335372-12463696
The Aesthetics of Bias in Machine-Mediated Speech Acts
Seval, H. D. (2026) · Springer eBooks · doi:10.1007/978-3-031-87993-7_186-1
Moganadas, S. R., Marín-González, F., Nun, S. H., & Gan, C. L. (2026) · Frontiers in Education · doi:10.3389/feduc.2026.1856839 · Cited by 1
Agha, Y., & Khan, M. S. A. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-8167-1.ch004
Towards Responsible Generation of Synthetic Datasets with Large Language Models
Franzoni, V. (2025) · IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) · doi:10.1109/wi-iat67162.2025.00123 · Cited by 1
Çelik, M. A., Duran, S., Duran, V., & Sevinç, H. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-1413-4.ch003
Tool or Companion? Reframing Conversational AI to Prevent Psychological Harm
Novelli, A. M., Shergill, S. S., & Teixeira, A. S. (2026) · JMIR Mental Health · doi:10.2196/99354 · Cited by 1
Human-Centric Auditing of AI-Powered Generative Search: When Citations Diverge from Rankings
Cheng, C.-Y. (2026) · Information Systems Frontiers · doi:10.1007/s10796-026-10813-4
Generative AI in Software Testing Education: Opportunities and Challenges
Warasart, M., Rattanathepee, U., Keerin, P., Sangounsak, K., Sunthorn, W., & Dabthong, H. (2026) · National Conference on Technical Education and International Conference on Technical Education (NCTechEd & ICTechEd) · doi:10.1109/icteched69022.2026.11682881
Oladunmoye, E. O., Adewusi, M. A., & Oyedele, L. O. (2026) · WAMDEVIN International Journal of Management Development and Computing · doi:10.68124/wamdevin-ijmdc/10092601
Karakaş, A., & Boonsuk, Y. (2026) · Innovation in Language Learning and Teaching · doi:10.1080/17501229.2026.2729892
Gil, W., & Sung, C. (2026) · Religions · doi:10.3390/rel17091069
Vision-Language Models on Edge Devices: Architectures, Optimization, and Real-World Deployment
Omar, K., & Gómez, J. M. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-0136-3.ch003
Breaking the shackles of confidentiality: Why AI should “free” scientific knowledge
Durmuş, H. (2026) · Research Ethics · doi:10.1177/17470161261488065
Human-AI Collaboration in Psychological Practice: A Relational Framework for Responsible AI
Ala, S. M. F., Galvão, A. M., & Lourenço, D. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9109-0.ch007
Zhang, C. (2026) · Open Computer Science · doi:10.1515/comp-2025-0073
“These concerns also reflect broader warnings about the risks of overconfident and ungoverned language model outputs in high-stakes real-world applications [5].”
Huang, K., Liu, C., & Yang, J. (2026) · Journal of Intelligence · doi:10.3390/jintelligence14090218
“LLMs are trained on corpora that include much of the public internet, which may contain the very creativity tasks, benchmarks, and scoring rubrics used to evaluate them (McCoy et al., 2023; Bender et al., 2021).”
InfluentialJustice by humans, assisted by AI: What should the future of adjudication look like?
Khan, A., Haider, A., & Salim, A. (2026) · Oñati Socio-legal Series · doi:10.35295/osls.iisl.2756 · Cited by 4
“Bender et al. (2021) describe large language models as “stochastic parrots”: systems that display a high degree of linguistic proficiency without possessing any model of the world, including the institutional world in which legal language operates.”
The question of AI authorship: Literary-theoretical perspectives
Duman, Ç. Ö., van Heerden, I., & Bas, A. (2026) · The Information Society · doi:10.1080/01972243.2026.2724248 · Cited by 1
Ng, L. H. X., & Chan, L. Q. (2026) · Data · doi:10.3390/data11090235
Possler, D., Dietrich, F., Scheper, J., Lammers, A., & Spatzenegger, A. (2026) · Publizistik · doi:10.1007/s11616-026-00942-w
Transcending natural evolution in the age of artificial intelligence
Guo, P., Jia, X., Tao, L.-Q., Yang, Y., & Ren, T.-L. (2026) · Science China Information Sciences · doi:10.1007/s11432-026-5041-7
“Many systems still struggle with systematic reasoning, counterfactual causal inference, and value-sensitive judgement [100, 101].”
Carruthers, Y. A., & Heim, J. (2026) · AI & Society · doi:10.1007/s00146-026-03365-3
Concept: Advancing Health Sciences Education Through an AI Companion Powering the Dynamic Classroom
Abdelbaset, F. (2026) · Springer eBooks · doi:10.1007/978-3-032-24819-0_15
Generative AI Technologies, Techniques and Tensions: A Primer
Behrens, J. T. (2026) · Springer eBooks · doi:10.1007/978-3-032-24819-0_2 · Cited by 2
Tang, X., Filonczuk, A., Zhang, X., & Cheng, Y. (2026) · Springer eBooks · doi:10.1007/978-3-032-24819-0_19
Seçkin, C. A. N. (2026) · Educational Assessment Evaluation and Accountability · doi:10.1007/s11092-026-09510-2
Labour Market Implications of Generative AI in an Emerging Economy: The Case of Türkiye
Yüncüler, B. G., Beyhan, B., & Akçomak, İ. S. (2026) · Science Technology and Society · doi:10.1177/09717218261480259
AI Literacies for Adults in a GenAI Intensive Workforce: A Thematic Literature Review (2022–2026)
Creely, E. (2026) · Journal of Adolescent & Adult Literacy · doi:10.1002/jaal.70075
Alviz-Meza, A., Valencia-Arias, A., Rojas-Flores, S., & Díaz, F. (2026) · Processes · doi:10.3390/pr14182917 · Cited by 1
Klapproth, F. (2026) · Education Sciences · doi:10.3390/educsci16091507 · Cited by 2
Algorithmic Bias and Logical Inconsistency in LLM-Generated Code
Pandeya, P. R., & Crk, I. (2026) · Big Data and Cognitive Computing · doi:10.3390/bdcc10090313
Dynamic Accent Accommodation: Training Learners for World Englishes using Generative Speech
Bayaro, M., Abdullahi, A. M., & Olorunfemi, P. A. (2026) · Academos Journal · doi:10.65921/ny1hcf55
Which Words to Teach: Word Selection in the Age of Large Language Models
Hiebert, E. H. (2026) · Encyclopedia · doi:10.3390/encyclopedia6090201
“An LLM learns from the form of language—the patterns in how words are used—not from the world those words describe [64]; its fluency with distributional structure can be mistaken for a grasp of meaning.”
Who designs the loop? Anthropology and the cybernetics of human–AI relations
Artz, M., & Koycheva, L. (2026) · Human Organization · doi:10.1080/00187259.2026.2715691 · Cited by 1
Nazhand, N. (2026) · Education Sciences · doi:10.3390/educsci16091504
Mihal, J., & Black, C. F. (2026) · Cambridge University Press eBooks · doi:10.1017/9781009862530
Educating minds with generative AI
Paolo, L. D. D., Clark, A., & Wachter, T. (2026) · Communications Psychology · doi:10.1038/s44271-026-00522-8 · Cited by 1
Li, H., Yeung, S. S., Xu, G., et al. (2026) · AI Brain and Child · doi:10.1007/s44436-026-00044-4
“As the experts pointed out, AI operates on probabilistic logic; that is, it retrieves and presents existing knowledge based on statistical like - lihoods (Bender et al., 2021).”
Qi, T. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1929537
“These concerns are sharper in physical training because inaccurate or overconfident wording can encourage compensatory movement or unsafe range (Bender et al., 2021; Ji et al., 2023).”
Onat, O. (2026) · AI and Ethics · doi:10.1007/s43681-026-01378-8
“Bender and colleagues [ 11 ] showed that models trained on vast, uncurated corpora inherit the associations and blind spots of those corpora at scale: the worldview of the data becomes the geometry of the space.”
Ruiu, M. L., Ruiu, G., & Ragnedda, M. (2026) · International Sociology · doi:10.1177/02685809261480038
Autissier, D., Michel, C., Gouesse, C., & Pouzeratte, F. (2026) · Question(s) de management · doi:10.3917/qdm.240.0049
Cultural bias or universal traits? Exploring personality-like profiles of large language models
Hang, Y., Soto, C. J., Lignier, B., Ling, Z., Liu, Q., Wang, Y., Zhu, J., He, X., & Wu, W. (2026) · Royal Society Open Science · doi:10.1098/rsos.252491
A Risk-Governed Deployment Framework for LLM-Powered Educational Companion Robots in K--12 Education
Wu, L., Zhang, M., Zhai, P., & Cao, Y. (2026) · International Conference on Digital Society and Artificial Intelligence (DSAI) · doi:10.1145/3833541.3833774
Devotional AI: digital companions and bad faith love
Krueger, J. W., & Osler, L. (2026) · Phenomenology and the Cognitive Sciences · doi:10.1007/s11097-026-10211-w
Should the Legal Concept of Speech Include Chatbot Output in US Law?
Dobler, T., & Bracker, D. (2026) · Philosophy & Technology · doi:10.1007/s13347-026-01187-1
“…contemporary AI systems lack the properties required for genuine linguistic agency, including concepts, beliefs, communicative intentions, authorship, and participation in human normative practices (Bender & Koller, 2020; Bender et al., 2021; Katzir, 2023; Shanahan, 2024; Murphy et al., 2025 ).”
Rivera, C., Tusa, F., Espinoza, M., & Aguaded, I. (2026) · Anàlisi · doi:10.5565/rev/analisi.3895
Sinanoğlu, N., Ismayilova, A., Muradova, A., Hajiyeva, R., Ahmadov, N., Hajiyev, Y., & Aliyeva, A. (2026) · European Archives of Oto-Rhino-Laryngology · doi:10.1007/s00405-026-10607-z · Cited by 1
Will AI save ethics—and make ethicists obsolete?
Tretter, M. (2026) · AI and Ethics · doi:10.1007/s43681-026-01384-w
Cavaleri, P. (2026) · JLIS it · doi:10.36253/jlis.it-749
“The result is a form of decontextualized or weakly contextualized knowledge: excerpts separated from the interpretive contexts that actually give them meaning, excerpts that create confusion and misinformation (Floridi 2020; Bender et al. 2021; Carr 2025).”
Generative AI and Foundation Models: Governance, Ethics, and Democratic Implications
Gupta, D., Anvar, P., Sаpаyev, V., Akhmedjon, Y., Mamurjon, O., Kamola, Z., Iskandarovich, R. O., & Nazirjon, R. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9164-9.ch003
Sreedhar, M. S., Vaidehi, V., & Patil, A. (2026) · International Conference on Trends in Quantum Computing and Emerging Business Technologies (TQCEBT) · doi:10.1109/tqcebt67648.2026.11681227
“Bender et al. [3] discussed the social and environmental risks of large language models, particularly their disproportionate impact on marginalized communities.”
Zhong, Y., El-Diraby, T. E., & Goodfellow, S. D. (2026) · Journal of Computing in Civil Engineering · doi:10.1061/jccee5.cpeng-7192
Qi, Z., Zhang, R., & Feng, Y. (2026) · Electronics · doi:10.3390/electronics15184212
Chen, J., Shu, Y., & Li, H. (2026) · Processes · doi:10.3390/pr14182954 · Cited by 1
Quem controla o passado? IA generativa e as disputas pela memória coletiva na era digital
Huertas, E. M. C. G., & Fernandes, C. M. (2026) · Anagramas - Rumbos y sentidos de la comunicación · doi:10.22395/angr.v25n49a11
“Tais modelos não compreendem semanticamente o mundo, mas operam sobre correlações estatísticas entre formas linguísticas e visuais, produzindo respostas plausíveis porque aprendem regularidades nos dados, e não porque possuem entendimento (Bender e Koller, 2020; Bender et al ., 2021).”
A Trust-Centric Evaluation Framework for RAG- Based Question Answering Systems
Karthikeyan, T., Devi, A., Kanishk, P. G., Karthick, G., & Kumar, C. M. (2026) · International Conference on Trends in Quantum Computing and Emerging Business Technologies (TQCEBT) · doi:10.1109/tqcebt67648.2026.11681692
Towards Bias-Aware and Responsible AI-Assisted Requirements Engineering
Cao, F., Hassan, S., Alwidian, S., & Perry, B. (2026) · IEEE International Requirements Engineering Conference Workshops (REW) · doi:10.1109/rew72749.2026.00044
“Large Language Models trained on publicly available texts and audio data often reflect dominant cultural and linguistic perspectives, which can marginalize minority groups during requirements elicitation [5].”
Lindebaum, D. (2026) · Academy of Management Learning and Education · doi:10.5465/amle.2026.0628 · Cited by 1
Natural Language Processing and Law
Cereceda, P. M. (2026) · Springer eBooks · doi:10.1007/978-3-031-87993-7_184-1
Reyes, A. T. C., Sánchez, M. A. F., Zúñiga, P. I. V., Palacios, I. A. M., & Palacios, G. M. M. (2026) · Ciencia Interdisciplinaria Internacional · doi:10.70577/cieninter.v4i3.131
Epistemic Injustice in LLMs: Mind Audit and Intervention Based on Thought Attribution
Zhai, Y. (2026) · IFIP advances in information and communication technology · doi:10.1007/978-3-032-37199-7_4
Saha, S., & Singh, M. K. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9089-5.ch001
Verhulsdonck, G., & Jiang, J. (2026) · Journal of Business and Technical Communication · doi:10.1177/10506519261485276
Landers, R. N. (2026) · The Journal of Applied Behavioral Science · doi:10.1177/00218863261488831
Interaction-induced knowledge narrowing risk in LLM systems
Rutherford, D., & Wu, N. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02101-6
“…shift selection pressure upstream into retrieval, ranking, and compression processes, consistent with broader critiques of what scaling does and does not resolve in large language systems [5], and with infrastructure-studies accounts of how capacity constraints relo-cate rather than disappear [28].”
ARTIFICIAL INTELLIGENCE (AI): BOON AND CURSE FOR GEN ALPHA AND BETA EDUCATION
Thirumoorthy, G., Datt, G., & Sathish, K. (2026) · International Journal of Research -GRANTHAALAYAH · doi:10.29121/granthaalayah.v14.i8.2026.7131
CHABI, O. M., Dadjo, S. D.-D. Y., & Hakibou, A. (2026) · Akofena revue scientifique des Sciences du Langage Lettres Langues & de la Communication · doi:10.48734/akofena.n022.vol.5.25.2026
Ideology by Alphabet: Option Order and the Measurement of Machine Political PreferencesPreprint
Olah, T., Erdey, L., Tőkés, T., & Nádasi, L. (2026) · Research Square · doi:10.21203/rs.3.rs-11063104/v1
do Nascimento, L. S. O., Abdalla, G. H. B., Pereira, J. A. D., et al. (2026) · Revista Master Ensino Pesquisa e Extensão · doi:10.47224/revistamaster.v10i20.757
Associations of Generative AI Use for Health Issues With Anxiety and Depression
Zhai, Y., Song, J., Jun, E., Lemberger‐Truelove, M. E., Lin, Y.-W. D., Xiong, Y., Kim, I., Shannonhouse, L., & Du, X. (2026) · Journal of Counseling & Development · doi:10.1002/jcad.70064 · Cited by 1
Teaching Craft Intelligence: Sovereignty Protocols for AI‐Mediated Student Leadership
Haskell, C. (2026) · New Directions for Student Leadership · doi:10.1002/yd.70070
“This increases the public realm (and the student leaders who inhabit it), vulnerable to manipulation and the erosion of a shared reality, especially when fluent model outputs are mistaken for knowledge rather than for pattern-completion over web-scale corpora (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.