Stochastic parrot

In machine learning, a "stochastic parrot" is[1] a large language model that is good at generating
convincing language, but does not actually understand the meaning of the language it is processing.[2] The
term was coined by Emily M. Bender[2][3] in the 2021 artificial intelligence research paper "On the
Dangers of Stochastic Parrots: Can Language Models Be Too Big?" by Bender, Timnit Gebru,
Angelina McMillan-Major, and Margaret Mitchell.[4]

Definition and implications

According to Lindholm, et. al., the analogy highlights two vital limitations:[1]

(i) The predictions made by a learning machine are essentially repeating back the contents of
the data, with some added noise (or stochasticity) caused by the limitations of the model.

(ii) The machine learning algorithm does not understand the problem it has learnt. It can't
know when it is repeating something incorrect, out of context, or socially inappropriate.

They go on to note that because of these limitations, a learning machine might produce results which are
"dangerously wrong".[1]

Origin
The term was first used in the paper "On the Dangers of Stochastic Parrots: Can Language Models Be Too
Big?" by Bender, Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell (using the pseudonym
"Shmargaret Shmitchell").[4] The paper covered the risks of very large language models, regarding their
environmental and financial costs, inscrutability leading to unknown dangerous biases, the inability of the
models to understand the concepts underlying what they learn, and the potential for using them to deceive
people.[6] The paper and subsequent events resulted in Gebru and Mitchell losing their jobs at Google, and
a subsequent protest by Google employees.[7][8]

Subsequent usage
In July 2021, the Alan Turing Institute hosted a keynote and panel discussion on the paper.[9] As of May
2023, the paper has been cited in 1,529 publications.[10] The term has been used in publications in the
fields of law,[11] grammar,[12] narrative,[13] and humanities.[14] The authors continue to maintain their
concerns about the dangers of chatbots based on large language models, such as GPT-4.[15]

See also
1 the Road – AI-generated novel
Chinese room
Criticism of artificial neural networks
Criticism of deep learning
Criticism of Google
Cut-up technique
Infinite monkey theorem
Generative AI
List of important publications in computer science
Markov text
Stochastic parsing

References
1. Lindholm et al. 2022, pp. 322–3.
2. Uddin, Muhammad Saad (April 20, 2023). "Stochastic Parrots: A Novel Look at Large
Language Models and Their Limitations" (https://towardsai.net/p/machine-learning/stochasti
c-parrots-a-novel-look-at-large-language-models-and-their-limitations). Towards AI.
Retrieved 2023-05-12.
3. Weil, Elizabeth (March 1, 2023). "You Are Not a Parrot" (https://nymag.com/intelligencer/artic
le/ai-artificial-intelligence-chatbots-emily-m-bender.html). New York. Retrieved 2023-05-12.
4. Bender, Emily M.; Gebru, Timnit; McMillan-Major, Angelina; Shmitchell, Shmargaret (2021-
03-01). "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" (https://
doi.org/10.1145/3442188.3445922). Proceedings of the 2021 ACM Conference on Fairness,
Accountability, and Transparency. FAccT '21. New York, NY, USA: Association for
Computing Machinery: 610–623. doi:10.1145/3442188.3445922 (https://doi.org/10.1145%2
F3442188.3445922). ISBN 978-1-4503-8309-7. S2CID 232040593 (https://api.semanticsch
olar.org/CorpusID:232040593).
5. "Stochastic" (https://www.merriam-webster.com/dictionary/stochastic). Merriam-Webster.
Retrieved 2023-05-13.
6. Haoarchive, Karen (4 December 2020). "We read the paper that forced Timnit Gebru out of
Google. Here's what it says" (https://www.technologyreview.com/2020/12/04/1013294/googl
e-ai-ethics-research-paper-forced-out-timnit-gebru/). MIT Technology Review. Archived (http
s://web.archive.org/web/20211006233625/https://www.technologyreview.com/2020/12/04/10
13294/google-ai-ethics-research-paper-forced-out-timnit-gebru/) from the original on 6
October 2021. Retrieved 19 January 2022.
7. Lyons, Kim (5 December 2020). "Timnit Gebru's actual paper may explain why Google
ejected her" (https://www.theverge.com/2020/12/5/22155985/paper-timnit-gebru-fired-google
-large-language-models-search-ai). The Verge.
8. Taylor, Paul (2021-02-12). "Stochastic Parrots" (https://www.lrb.co.uk/blog/2021/february/sto
chastic-parrots). London Review of Books. Retrieved 2023-05-09.
 9. Weller (2021).
10. "Bender: On the Dangers of Stochastic Parrots" (https://scholar.google.com/scholar?cluster=
415069420329958137). Google Scholar. Retrieved 2023-05-12.
11. Arnaudo, Luca (April 20, 2023). "Artificial Intelligence, Capabilities, Liabilities: Interactions in
the Shadows of Regulation, Antitrust – And Family Law". SSRN. doi:10.2139/ssrn.4424363
(https://doi.org/10.2139%2Fssrn.4424363). S2CID 258636427 (https://api.semanticscholar.o
rg/CorpusID:258636427).
12. Bleackley, Pete; BLOOM (2023). "In the Cage with the Stochastic Parrot" (https://specgram.c
om/CXCII.3/07.bloom.cage.html). Speculative Grammarian. CXCII (3). Retrieved
2023-05-13.
13. Gáti, Daniella (2023). "Theorizing Mathematical Narrative through Machine Learning".
Journal of Narrative Theory. Project MUSE. 53 (1): 139–165. doi:10.1353/jnt.2023.0003 (http
s://doi.org/10.1353%2Fjnt.2023.0003). S2CID 257207529 (https://api.semanticscholar.org/C
orpusID:257207529).
14. Rees, Tobias (2022). "Non-Human Words: On GPT-3 as a Philosophical Laboratory".
Daedalus. 151 (2): 168–82. doi:10.1162/daed_a_01908 (https://doi.org/10.1162%2Fdaed_a
_01908). JSTOR 48662034 (https://www.jstor.org/stable/48662034). S2CID 248377889 (http
s://api.semanticscholar.org/CorpusID:248377889).
15. Goldman, Sharon (March 20, 2023). "With GPT-4, dangers of 'Stochastic Parrots' remain,
say researchers. No wonder OpenAI CEO is a 'bit scared' " (https://venturebeat.com/ai/with-g
pt-4-dangers-of-stochastic-parrots-remain-say-researchers-no-wonder-openai-ceo-is-a-bit-sc
ared-the-ai-beat/). VentureBeat. Retrieved 2023-05-09.

Works cited
Lindholm, A.; Wahlström, N.; Lindsten, F.; Schön, T. B. (2022). Machine Learning: A First
Course for Engineers and Scientists. Cambridge University Press. ISBN 978-1108843607.
Weller, Adrian (July 13, 2021). On the Dangers of Stochastic Parrots: Can Language Models
Be Too Big? (https://www.youtube.com/watch?v=N5c2X8vhfBE) (video). Alan Turing
Institute. Keynote by Emily Bender. The presentation was followed by a panel discussion.

Further reading
Thompson, E. (2022). Escape from Model Land: How Mathematical Models Can Lead Us
Astray and What We Can Do about It. Basic Books. ISBN 978-1541600980.

External links
"On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? (https://common
s.wikimedia.org/wiki/File:On_the_Dangers_of_Stochastic_Parrots_Can_Language_Models
_Be_Too_Big.pdf)" at Wikimedia Commons

Retrieved from "https://en.wikipedia.org/w/index.php?title=Stochastic_parrot&oldid=1163666562"