- Can a language model feel pain?
Clearly not. Large language models are computer software and they can no more suffer than any other human-created artifact can.
Replying to
Remote
Prof. Emily M. Bender(she/her)
@emilymbender@dair-community.social
mastodon 4.7.3Professor, Linguistics, University of Washington
Faculty Director, Professional MS Program in Computational Linguistics (CLMS)
If we don't know each other, I probably won't reply to your DM. For more, see my contacting me page: http://faculty.washington.edu/ebender/contact/
29139 Followers
344 Following
50 Posts
Joined November 04, 2022
Boosted by @trending@homestead.social
Hearing lots of discourse about "stochastic parrots" recently? Here's a discussion of the backstory, both of the phrase itself and what's going on now.
https://www.bloodinthemachine.com/p/the-two-words-the-ai-industry-hates
Boosted by @trending@homestead.social
I've been talking to journalists a LOT recently, and it was particularly delightful to get to speak about questions of linguistics, language change and prescriptivism, for once:
https://www.cnet.com/tech/services-and-software/ai-not-super-intelligence-no-one-should-use-trumps-term/
Boosted by @MaryAustinBooks@mstdn.social
New from @timnitGebru@dair-community.social and me:
"It is high time for us to regulate usage of systems sold as “AI”, not because they are magical machines, but because they are error-prone products that shouldn’t be used in high stakes scenarios."
https://www.theguardian.com/commentisfree/2026/oct/01/forget-superintelligence-error-prone-ai-nearly-sparked-world-war-iii-this-month
Boosted by @trending@homestead.social
Have you seen Ghost in the Machine yet? If not, you are in for a treat --- a thoughtful, wry deep dive into the history, present, and ideology of "AI". Streaming for free on PBS's YouTube channel in the US & Korea, rentable on Kinema elsewhere.
https://notaidoc.com/
I appreciated this chance to go on Bloomberg Tech, though it's always a bit odd to have a conversation with a reporter who is really interested in what's effectively a fictional story.
https://www.bloomberg.com/news/videos/2026-10-06/bender-ai-s-existential-risk-is-fake-video
Here's the thing about social media folks:
Post on your own feed = people who follow you and presumably want to know your thoughts will see it.
Reply to someone else's post = you are demanding attention from someone else, so you'd be well advised to check whether you are telling them something they likely already know.
Wanna share a thought based on what someone else said? Bare minimum, respect their expertise and acknowledge their contribution.
Boosted by @trending@homestead.social
RE: https://dair-community.social/@emilymbender/110906329010147244
Every time I decide to post a thread to Mastodon, I'm hit with a new wave of mansplaining. What is it with the culture here? Why do all of these guys think it's so important to share their deep insights in my mentions when they can just post on their own feed? FFS.
Quoting
I use Mastodon for public scholarship, because that is one of the main benefits I see to public/open social media. It's how I used Twitter and I've definitely benefited from other people using social media this way.
But in my experience, Mastodon is 'splainy AF, which is exhausting. Just this morning, a gentleman decided I would benefit from an explanation of GIGO, FFS.
So, while I never insist on titles, I'm going to include mine in my display name for a while, to see if that helps.
Open quoted postBoosted by @trending@homestead.social
Created this for a reply on LinkedIn, but it applies pretty generally, so posting it here, too:
Boosted by @trending@homestead.social
I think there are two key lessons from Linguistics as we deal with LLMs/chatbots being imposed on us everywhere. Grateful for Undark Magazine giving space for me to lay these out:
https://undark.org/2026/09/25/interview-artificial-intelligence/
Boosted by @trending@homestead.social
From @timnitGebru@dair-community.social and me on what we should take from the 2026 summer of AI hype:
https://www.technologyreview.com/2026/09/22/1144867/dont-be-fooled-summer-ai-hype/
Boosted by @trending@homestead.social
Me: I do not need the mansplaining mansplained. No idea why you thought that might be welcome.
Mansplainer: Buh buh buh why are you so rude to meeeee
---
Me: Replying to someone is demanding their attention.
Mansplainer: Well actually people have different mental models of social media.
---
Starting to see the pattern: mansplainers don't understand consent, but feel VERY ENTITLED to push their uninteresting thoughts into other people's mentions.
Boosted by @trending@homestead.social
All of this coverage of AI Doomerism making you anxious? It's nonsense and you can go back to your more pressing concerns. If you want to learn more:
https://buttondown.com/maiht3k/archive/what-to-read-to-stay-grounded-amidst-ai-doomerism/
Boosted by @trending@homestead.social
Thinking about the selling point of "AI" supposedly "freeing up time" to do important things. Beyond devaluing the tasks that we can supposedly offload to software systems, it occurs to me that this is kind of opposite of mindfulness/being in the moment.
A short thread >>
Me: Puts time and effort into curating a reading list (nb: this is not my day job) to help people with navigating this bananas moment where the media is captured by all the worst AI bros, posts link here on the Fedi.
Mansplainer in my replies on Mastodon: "It's just marketing"
Me: YES, I KNOW. Also, do you realize it's possible to either keep your thoughts to your self or just post on your own feed? Sheesh.
Replying to
They can't, because the Doomer narrative *is* the Booster narrative. To debunk the Doomer stuff (and it *is* bunk), you have to point out that it's based on pure fantasy that the mathy maths are going to combust into consciousness ...
... or if not consciousness, then some thinking, reasoning, self-improving entity.
>>
Replying to
The Boosters say: AI is a thing, it's inevitable, it's imminent and it's going to solve all of our problems.
The Doomers say: AI is a thing, it's inevitable, it's imminent and it's going to kill us all.
These are obviously the same story, with just different twists at the end. And both are nonsense. There is "moderate" position between them, but there is space outside of that fantasy world where we can see the actual harms happening in the name of "AI".
>>
Boosted by @trending@homestead.social
I'm afraid the Doomer discourse has persisted for another week (UN "AI" briefing, etc) so in case you missed this when we first put it out earlier in this bananas news cycle, I'm reposting it:
https://buttondown.com/maiht3k/archive/what-to-read-to-stay-grounded-amidst-ai-doomerism/
w/ @alex@dair-community.social and @deccamuldowney@journa.host
Boosted by @oxy@social.bsdlab.au
RE: https://dair-community.social/@emilymbender/117271327519433700
Folks, if you want to know what's *really* up with "AI", where the harms lie, and contextualize the nonsense du jour, this film is must watch.
Quoting
For audiences in the US and Korea, Ghost in the Machine is now available on PBS's YouTube!
https://www.youtube.com/watch?v=dLQzVLtvdv8
Replying to
I want everyone, especially journalists, to notice the silence of the tech execs about this. This includes people like Bill Gates and Mark Zuckerberg whose recent blog posts talk about balancing the benefits and harms of "AI", and don't include Doomer scenarios among the harms.
They are complicit in the misinforming of the public. They are complicit in stoking everyone's fear. They reap the benefits of the Doomer narrative because it is their narrative.
/fin
On the next Mystery AI Hype Theater 3000 livestream, we’re so excited to chat with authors of the new book, Against Tech Oligarchy! JS Tan]and Clarissa Redwine will join me and @alex@dair-community.social to chat about tech industry labor. Tune in:
Monday, Oct 5, noon Pacific
https://twitch.tv/dair_institute
For audiences in the US and Korea, Ghost in the Machine is now available on PBS's YouTube!
https://www.youtube.com/watch?v=dLQzVLtvdv8
Ghost in the Machine | Full Documentary
Boosted by @oxy@social.bsdlab.au
Quoting
On the next Mystery AI Hype Theater 3000 livestream, we’re so excited to chat with authors of the new book, Against Tech Oligarchy! JS Tan]and Clarissa Redwine will join me and @alex@dair-community.social to chat about tech industry labor. Tune in:
Monday, Oct 5, noon Pacific
https://twitch.tv/dair_institute
Open quoted postReplying to
But please pause for a moment and think about what metaphors "unleash" rests on. What do we put leashes on? (Mostly dogs) Why do we keep dogs on leashes? (To keep them from running off and hurting themselves, other animals, or people.)
So when you talk about "unleashing" a software system, what properties does that framing attribute to the system?
>>
Replying to
As @alex@dair-community.social and I write in The AI Con (see cp 6): even though the Boosters and the Doomers like to portray themselves as holding down opposite extremes of opinions on "AI", they are two sides of the same coin.
https://thecon.ai
>>
Replying to
As I was walking home yesterday, I got flagged down by my neighbor who had heard a panic-inducing story about rogue AI from a friend over lunch and hoped I could talk her down.
I could and I did! But I am recognizing her distress (likely repeated across many many ordinary people) as yet another harm perpetuated by the people selling "AI" narratives.
>>
Boosted by @trending@homestead.social
RE: https://newsie.social/@TheNeedling/117290024708643957
We effectively did a whole podcast on this, but the Needling gets the point across in a single short piece:
https://www.buzzsprout.com/2126417/episodes/19817583-manifesting-the-end-of-tech-bro-manifestos-2026-09-08
Quoting
New AI Literacy Program Awards Students Free Pizza If They Don’t Know How to Read Books Anymore: https://theneedling.com/2026/09/17/new-ai-literacy-program-awards-students-free-pizza-if-they-dont-know-how-to-read-books-anymore/
Replying to
Nadella, Pichai, Huang, Zuckerberg, Huang, Dean --- you all own this unhinged narrative just as much as the Doomers do.
Of course, they're happy to own this, because they benefit as much as Altman, Amodei, and Musk from keeping the policymakers focused on fantasies.
Replying to
Where are Satya Nadella, Sundar Pichai, Mark Zuckerberg, Jensen Huang, XD Huang, Jeff Dean on this? Why aren't they out there calling out the Doomer narrative for the nonsense that it is?
I have an hypothesis:
>>
Mystery AI Hype Theater 3000, Episode 85:
https://www.buzzsprout.com/2126417/episodes/19817583-manifesting-the-end-of-tech-bro-manifestos-2026-09-08
… in which @alex@dair-community.social and I sift through the ramblings of billionaires who understand neither the world we live in nor the tech they’re selling.
Boosted by @trending@homestead.social
Happy UK paperback release day to our book!
https://www.penguin.co.uk/books/468070/the-ai-con-by-hanna-emily-m-bender-and-alex/9781529949902
w/ @alex@dair-community.social
From Altman proclaiming the “singularity”, to the Guardian publishing blather about possible AI consciousness, to OpenAI and Anthropic blaming their cybercrimes on “rogue models” the nonsense is coming fast and thick these days and much of the media is lapping it up. Time for another Mystery AI Hype Theater 3000 livestream with me and @alex@dair-community.social:
Monday, August 3, noon PT
https://twitch.tv/dair_institute
RE: https://dair-community.social/@emilymbender/117292656176293011
Also available as video on PeerTube:
https://peertube.dair-institute.org/w/b8fH3fp2fr3Dpki11tsWfb
Quoting
Mystery AI Hype Theater 3000, Episode 85:
https://www.buzzsprout.com/2126417/episodes/19817583-manifesting-the-end-of-tech-bro-manifestos-2026-09-08
… in which @alex@dair-community.social and I sift through the ramblings of billionaires who understand neither the world we live in nor the tech they’re selling.
A couple of days ago I did a jetlagged live posting on bsky about the terrible op-ed in the Guardian about "AI" consciousness. Not going to port it bit by bit over here, but someone used skywriter to put it all in one place.
For your enjoyment, some ridicule as praxis:
https://skywriter.blue/@emilymbender.bsky.social/3mqzj7xzqfc2i
LinkedIn where people come to proudly state that they have cooked their own brains and/or just really really love licking those boots.
All of these were in response to my post about the ACM's announcement that they've decided "Now is the Time to Give LLMS Access to the ACM Digital Library" and then requestd input on that decision.
Replying to
Agreed that AI slop is a problem for scientific publishing (and the information ecosystem at large). But what's missing here is any mention of consequences for those who produce and submit it.
"AI" errors don't just appear in papers. They are there because of specific decisions, most proximally the decision by authors (or maybe "authors") to use extruded academic-looking text in their submissions.
>>
Quoting
From Altman proclaiming the “singularity”, to the Guardian publishing blather about possible AI consciousness, to OpenAI and Anthropic blaming their cybercrimes on “rogue models” the nonsense is coming fast and thick these days and much of the media is lapping it up. Time for another Mystery AI Hype Theater 3000 livestream with me and @alex@dair-community.social:
Monday, August 3, noon PT
https://twitch.tv/dair_institute
Open quoted postReplying to
LinkedIn posts (see prev toot for context)
One of my worst speech technology nightmares has been happening right in my own city! In the latest Mystery AI Hype Theater 3000, @alex@dair-community.social and I dig into it:
https://www.buzzsprout.com/2126417/episodes/19535278-state-of-emergency-2026-06-29
Replying to
Also, although synthetic text is never a good idea, this particular type of synthetic text (the sycophantic type) results from specific design decisions rather than any inherent properties of LLMs.
>>
RE: https://dair-community.social/@emilymbender/116974840056295473
Also available as video on PeerTube:
https://peertube.dair-institute.org/w/23g3vAkBsgQM1svfo3hhRD
Replying to
This next paragraph also betrays confusion on the part of the author as to what they are describing. AlphaFold and GNoME (called LLMs above) aren't chatbots and so don't provide the kind of output that is called or experienced as "sycophantic".
>>
Replying to
Replying to
@nefles@piaille.fr @alex@dair-community.social Not out, and also so far no publishers have been interested, alas.
Spanish and Italian are out, Russian, Chinese, Korean, Japanese and Turkish are on the way.
Replying to
@Npars01@mstdn.social @alex@dair-community.social thank you!