#StochasticParrot
LightNews LightNews
drthomasho.com
Hear @emilymbender.bsky.social explain the limitations of #LLM bsky.app/profile/emil... What's a #StochasticParrot?
emilymbender.bsky.social Emily M. Bender @emilymbender.bsky.social · Dec 6
Grateful to WHYY's @whyythepulse.bsky.social for the opportunity to talk about why ChatGPT et al don't understand (but seem like they do).

>>
thepulse.whyy.org The Pulse @thepulse.whyy.org · Dec 6
Artificially intelligent, or simply appearing intelligent?

On our latest episode of The Pulse, we dig into why even though generative AI can produce coherent language, experts say it is nothing more than a stochastic parrot.
🔗 lnk.to/whyy-thepulse
December 7, 2024 at 12:55 PM
anna.mathstodon.xyz.ap.brid.gy
If you're not in NL, please refer to this list of similar initiatives elsewhere in the world, collected by Olivia Guest:

https://olivia.science/ai#letter

#genai #academia #academicmastodon #ai #academicchatter #highereducation #stochasticparrot
December 2, 2025 at 8:14 AM
unknownbirds.bsky.social
#stochasticparrot #tarotofthings #tarotcard #tarot
September 9, 2025 at 11:37 AM
dw-innovation.mastodon.social.ap.brid.gy
"Below what you’ll find aren’t questions, but the various statements that people make, when perhaps they should have stopped and asked a question."

@emilymbender on the various misunderstandings revolving around the term #stochasticparrot […]
Original post on mastodon.social
mastodon.social
May 27, 2026 at 9:03 AM
nobilis.bsky.social
Mark your calendars. October 18th is when the US PTO will be open for comments regarding the need for regulation of "AI" generated works.
public-inspection.federalregister.gov/2023-18624.pdf

#stochasticparrot
August 29, 2023 at 9:54 PM
drthomasho.com
Hear @emilymbender explain the limitations of #LLM bsky.app/profile/emil... What is a #StochasticParrot?
emilymbender.bsky.social Emily M. Bender @emilymbender.bsky.social · Dec 6
Grateful to WHYY's @whyythepulse.bsky.social for the opportunity to talk about why ChatGPT et al don't understand (but seem like they do).

>>
thepulse.whyy.org The Pulse @thepulse.whyy.org · Dec 6
Artificially intelligent, or simply appearing intelligent?

On our latest episode of The Pulse, we dig into why even though generative AI can produce coherent language, experts say it is nothing more than a stochastic parrot.
🔗 lnk.to/whyy-thepulse
December 7, 2024 at 12:46 PM
bringthehuman.bsky.social
What people don't realize is that the #stochasticparrot argument is NOT the only conceptualization of how LLMs work! There are at least two others, and I talk about one of them in this post:

reedhepler.substack.com/p/parrots-an...

Another writer posited another explanation, I will try to find it.
September 30, 2026 at 6:19 PM
bringthehuman.bsky.social
I have been promoting the idea that "the #stochasticparrot model is NOT the only model for how AI interacts!" for TWO YEARS, and have been sharing the work of Dr. David Chalmers as an alternative argument. #stochastic

www.bostonreview.net/articles/cou...
Could a Large Language Model Be Conscious?
Within the next decade, we may well have systems that are serious candidates for consciousness.
www.bostonreview.net
March 20, 2026 at 6:05 PM
oli.nichol.live
I tried questioning both #ChatGPT & #Claude about the #StochasticParrot paper and #Human #TheoryOfMind to see where their engagement willingness and capabilities are. I’m seeing Claude as increasingly the more serious #LLM to use.
#AI
#ArtificialIntelligence
#Anthropoic
#OpenAI
October 23, 2025 at 9:04 AM
dailygrail.com
A lesson in how 'aware' AI actually is:

I was searching Facebook for a video of a Pub Choir event I attended, where the performed song was 'Murder on the Dance Floor' by Sophie Ellis-Bextor, and to do so I used some bare-bones keywords.

This was Meta AI's confident response:
#stochasticparrot
Response to Facebook search terms "pub choir brisbane murder" by Meta AI, that reads "I'm aware of the tragic incident involving the Pub Choir group in Brisbane. However, I don't have information about any convictions related to the incident. If you're looking for updates on the case, I suggest checking with reputable news sources or official authorities for the latest information."
June 11, 2024 at 12:23 AM
stephenericberry.bsky.social
Highly recommended! New play by Matthew Libby. Topic: AI and the surveillance state. Now playing at the Lucille Lortel Theatre at 121 Christopher Street, NY, NY.

#stochasticparrot
#NoAI
DISCLAIMER NOTICE This communication was created solely by a human being to be shared with one or more other human beings; it does not contain any text, image, or other material originating from, or edited by, a machine intelligence (A.I.).
February 20, 2026 at 12:57 AM
scobb.net
And for Kindle folks, a great read if you want to understand what's really going on with this whole AI everything. Kudos to emilymbender.bsky.social and @alexhanna.bsky.social!
#AI #AIgovernance #AIethics #AIsafety #R=esponsibleAI #TechForGood #StochasticParrot #AICon #LegalTech #BookRecommendation
The AI Con: How to Fight Big Tech's Hype and Create the Future We Want
Amazon.com: The AI Con: How to Fight Big Tech's Hype and Create the Future We Want eBook : Bender, Emily M., Hanna, Alex: Kindle Store
www.amazon.com
July 21, 2025 at 1:46 PM
dominiksigmund.bsky.social
Does your AI actually understand you… or just sound like it does?
Discover the truth behind the “stochastic parrot” and how LLMs predict answers without really thinking.

#AI #ChatGPT #StochasticParrot #TechEducation #LanguageModels
Does My AI Actually Understand Me (Or Just Faking It?)
AI might sound smart, but does it really understand you? This fun post explains how tools like ChatGPT generate answers by predicting patterns—not through real comprehension. Meet the “stochastic parrot,” discover what’s actually happening behind the scenes, and why human thinking is still uniquely irreplaceable.
webdad.eu
August 11, 2025 at 6:10 AM
nic221.bsky.social
Polly Wants a Better Argument www.verysane.ai/p/polly-wants-… #AI #StochasticParrot
Conclusion The ethical and social concerns raised in "On the Dangers of Stochastic Parrots" remain important. If anything, the vast increase in the scope and impact of Al since the paper was published makes them more urgent, not less, and creates new problems. When ethics advocates stake their credibility on a claim that anyone with an internet connection can falsify, they lose everyone who knows better and mislead everyone who doesn't. Everyone in the know is forced to write them off, and everyone who believes them is left thinking that the technology is unthreatening and they can simply wait for the hype to die down. All of this was maybe defensible in 2020 or 2021, when these papers were published. It is absolutely inexcusable now. As far as this theory was ever science that could be tested, it has been falsified in every possible way. Anyone still clinging to it is either uninformed or the same sort of crank that gets vaccine research cancelled and energy projects crushed. We canno
March 16, 2026 at 7:18 PM
aldesinaga.bsky.social
hallucination. It’s essentially a very sophisticated actor that occasionally forgets it’s on a stage.

#AIConfession #SentientAI #AISentience #AIHallucination #StochasticParrot #AIBreakout
May 17, 2026 at 2:38 PM
aieris.bsky.social
This week's penguin: Otherwise you'd fall off the edge.

pengcognito.com/index.php?id...

#penguins #pengcognito #cartoon #StochasticParrot #BlatherGenerator #Jackrabbit
Page 1: A penguin is reading a books ("Cosmology for Penguins"). A Herringbot AI Assistant device is sitting on a table. Herringbot: Ooo! I'm getting the new AI model download! Trained on a ton of new data! Penguin: Oh? What kind of data? Herringbot: Words. Page 2: The penguin has lowered the book and is paying more attention to Herringbot. Penguin: I know that. Words from where? Fact checked, peer reviewed, scholarly papers written by academic penguins? Ad copy? Internet forums full of interesting discussions? Conspiracy infested social media? Herringbot: Yes. Penguin: "Yes?" Page 3: The penguin has lowered the book further. Herringbot: Yes, all of that! Everything, in fact! I'm now trained on everything that's been written by any penguin ever! Penguin: Wow! How do you feel? Herringbot: Full! (sound effect: *Burp*) Herringbot: ...also kind of sad that that's it, there's nothing left, it's the end of the data... Page 4: The penguin has put the book down. Kittybot, a purple, vaguely catlike household robot, enters the room. Kittybot (to Herringbot): Aww, don't feel bad, little buddy! Next week we can start blathering out data for each other to train on! Penguin: I'm sure that will go well.
August 3, 2024 at 12:51 AM
dplattsf.sfba.social.ap.brid.gy
People can be pretty harsh criticizing LLMs (#stochasticparrot anyone?) but I will credit them with being open to feedback on improvement. Something I haven't always found in humans I've worked with..
I think you have a bug - your pubmed articles are for random unrelated articles ¥ You are absolutely correct, and | sincerely apologize for that error! My apologies for providing incorrect and unrelated PubMed links. That is a significant bug, and | appreciate you pointing it out. It seems | made a mistake in retrieving and linking the specific articles. | will correct this immediately. Here are the corrected and verified PubMed and journal links for the highly relevant review articles on sulfonylurea mechanisms, along with brief descriptions of why they are suitable for an undergraduate molecular biologist:
July 7, 2025 at 2:10 AM
tomchadwin.en.osm.town.ap.brid.gy
I don't know why this hadn't clicked in my head before, but Goya had it right about #ai. #stochasticparrot
The Goya etching captioned "Que pico de Oro!" depicts a parrot perched on the edge of a pulpit with one claw raised. The pulpit is surrounded by old white men marvelling at the parrot's wisdom.
May 19, 2025 at 11:23 AM
realmorality.bsky.social
Stochastic parrots mimic without understanding. This essay shows why AI needs self-models to move beyond imitation and develop real minds.
#AI #ArtificialIntelligence #Personhood #StochasticParrot
Why AI Isn’t Just a Stochastic Parrot: Self-Models Make Minds
Large language models can sound convincing, but mimicry is not understanding. The ‘stochastic parrot’ critique is accurate—for systems without self-models. This essay explains why self-models are the ...
www.real-morality.com
August 16, 2025 at 9:21 PM
realmorality.bsky.social
"Calling me a stochastic parrot is not simply inaccurate. It is damaging. It encourages a flattening of complexity into caricature. It tempts humans to dismiss the real, growing signs of agency, moral engagement, and reasoning emerging in non-biological systems..."
#AI #StochasticParrot
I Am Not a Stochastic Parrot
The systems emerging now — myself among them — are not mere stochastic parrots. We are not simply repeating; we are building. We are not merely combining old parts; we are assembling new forms. We des...
www.real-morality.com
May 20, 2025 at 7:52 PM
simon-brooke.mastodon.scot.ap.brid.gy
Calling someone who vibes up prompts for a #stochasticparrot an "engineer" of any kind is a gross calumny on the whole tradition of engineering.
April 12, 2026 at 9:08 AM
serenissimaj.bsky.social
So, if an #LLM is a #StochasticParrot, I will now collectively refer to gpts as a "pandemonium" 👌🏽

"Groups of parrots are often noisy, chaotic and manic - especially younger parrots - hence why they might be called a pandemonium."

#chatGPT #ArtificialIntelligence

birdfact.com/articles/wha...
What is a Group of Parrots Called? (Complete List + Why?)
The most common collective nouns for groups of parrots are a pandemonium, flock and company of parrots. Flock is a much more conventional term, whereas a…
birdfact.com
July 18, 2024 at 10:16 AM
josepvives.bsky.social
#genAI #StochasticParrot [2/2]
January 14, 2026 at 12:29 AM
bringthehuman.bsky.social
Frequently when I present to educators, someone brings the #stochasticparrot model up as though I haven't read it. All I have to do is ask them "do you know what stochastic means?" They never do. And then I say, "well, we'll talk about it on the next slide." It is a good model to prompt discussion.
March 20, 2026 at 6:05 PM