Emily M. Bender
@emilymbender.bsky.social
Reposted by Emily M. Bender
Yes. Yes. Yes.
Students use LLMs because they have never, not just now, had enough time & space for learning.
And they are frequently missing knowledge about how the various resources work. We just skipped right over that to "they're going to use it anyway so we have to integrate it into pedagogy."
Students use LLMs because they have never, not just now, had enough time & space for learning.
And they are frequently missing knowledge about how the various resources work. We just skipped right over that to "they're going to use it anyway so we have to integrate it into pedagogy."
In early 2024, I got to speak to Deans from up & down the West Coast of the US & Canada about "AI". I told them that the only value of ChatGPT for university administrators is as a contrast-dye test to show where resources are lacking.
December 9, 2025 at 2:36 PM
Yes. Yes. Yes.
Students use LLMs because they have never, not just now, had enough time & space for learning.
And they are frequently missing knowledge about how the various resources work. We just skipped right over that to "they're going to use it anyway so we have to integrate it into pedagogy."
Students use LLMs because they have never, not just now, had enough time & space for learning.
And they are frequently missing knowledge about how the various resources work. We just skipped right over that to "they're going to use it anyway so we have to integrate it into pedagogy."
It seems like the most likely thing here is that the voice was synthetic and it was reading off LLM generated strings that were lightly vetted/accepted by some call center worker somewhere. Terrible for everyone involved, really:
Had a really disconcerting interaction with what I'm pretty sure was an artificial customer service agent today.
Short thread >>
Short thread >>
December 9, 2025 at 1:54 PM
It seems like the most likely thing here is that the voice was synthetic and it was reading off LLM generated strings that were lightly vetted/accepted by some call center worker somewhere. Terrible for everyone involved, really:
Had a really disconcerting interaction with what I'm pretty sure was an artificial customer service agent today.
Short thread >>
Short thread >>
December 9, 2025 at 12:55 AM
Had a really disconcerting interaction with what I'm pretty sure was an artificial customer service agent today.
Short thread >>
Short thread >>
Reposted by Emily M. Bender
For everyone working on the intersections between linguistic and computational research, consider submitting to the upcoming SCiL! We're very excited that it will be co-located with ACL 2026 as a workshop, and we've also received NSF funding to help cover costs.
sites.google.com/view/scil2026
sites.google.com/view/scil2026
SCiL 2026
SCiL 2026 July 3rd and 4th San Diego, CA
sites.google.com
December 8, 2025 at 11:41 PM
For everyone working on the intersections between linguistic and computational research, consider submitting to the upcoming SCiL! We're very excited that it will be co-located with ACL 2026 as a workshop, and we've also received NSF funding to help cover costs.
sites.google.com/view/scil2026
sites.google.com/view/scil2026
Reposted by Emily M. Bender
🎧TODAY: Tune in to the livestream of the Mystery AI Hype Theater 3000 podcast 🎧
@emilymbender.bsky.social and @alexhanna.bsky.social will be digging into one recent attempt to define so-called "AGI"...
TODAY, December 8, noon Pacific
www.twitch.tv/dair_institute
@emilymbender.bsky.social and @alexhanna.bsky.social will be digging into one recent attempt to define so-called "AGI"...
TODAY, December 8, noon Pacific
www.twitch.tv/dair_institute
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
www.twitch.tv
December 8, 2025 at 5:10 PM
🎧TODAY: Tune in to the livestream of the Mystery AI Hype Theater 3000 podcast 🎧
@emilymbender.bsky.social and @alexhanna.bsky.social will be digging into one recent attempt to define so-called "AGI"...
TODAY, December 8, noon Pacific
www.twitch.tv/dair_institute
@emilymbender.bsky.social and @alexhanna.bsky.social will be digging into one recent attempt to define so-called "AGI"...
TODAY, December 8, noon Pacific
www.twitch.tv/dair_institute
Today!
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
twitch.tv
December 8, 2025 at 3:19 PM
Today!
Reposted by Emily M. Bender
This thread up and down.
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Here's the reality this example illustrates:
It's not even just about people blindly trusting what ChatGPT tells them. LLMs are poisoning the entire information ecosystem. You can't even necessarily trust that the citations in a published paper are real (or a search engine's descriptions of them).
It's not even just about people blindly trusting what ChatGPT tells them. LLMs are poisoning the entire information ecosystem. You can't even necessarily trust that the citations in a published paper are real (or a search engine's descriptions of them).
December 7, 2025 at 9:54 PM
This thread up and down.
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Reposted by Emily M. Bender
Scratch a definition of general intelligence, find a eugenicist. That and more on tomorrow's Mystery AI Hype Theater 3000!
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
twitch.tv
December 7, 2025 at 10:41 PM
Scratch a definition of general intelligence, find a eugenicist. That and more on tomorrow's Mystery AI Hype Theater 3000!
This thread up and down.
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Here's the reality this example illustrates:
It's not even just about people blindly trusting what ChatGPT tells them. LLMs are poisoning the entire information ecosystem. You can't even necessarily trust that the citations in a published paper are real (or a search engine's descriptions of them).
It's not even just about people blindly trusting what ChatGPT tells them. LLMs are poisoning the entire information ecosystem. You can't even necessarily trust that the citations in a published paper are real (or a search engine's descriptions of them).
December 7, 2025 at 9:54 PM
This thread up and down.
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Also, I've been beating this drum about the danger to the information ecosystem for six and half years now. A few links below:
Reposted by Emily M. Bender
Let me tell you a story. Perhaps you can guess where this is going... though it does have a bit of a twist.
I was poking around Google Scholar for publications about the relationship between chatbots and wellness. Oh how useful: a systematic literature review! Let's dig into the findings. 🧵
I was poking around Google Scholar for publications about the relationship between chatbots and wellness. Oh how useful: a systematic literature review! Let's dig into the findings. 🧵
December 5, 2025 at 10:35 PM
Let me tell you a story. Perhaps you can guess where this is going... though it does have a bit of a twist.
I was poking around Google Scholar for publications about the relationship between chatbots and wellness. Oh how useful: a systematic literature review! Let's dig into the findings. 🧵
I was poking around Google Scholar for publications about the relationship between chatbots and wellness. Oh how useful: a systematic literature review! Let's dig into the findings. 🧵
Join us tomorrow!
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
twitch.tv
December 7, 2025 at 2:13 PM
Join us tomorrow!
Reposted by Emily M. Bender
December 6, 2025 at 10:24 PM
Reposted by Emily M. Bender
Mind you, the other big Meta bet is making is "Artificial Superintelligence" -- being led by a 28 year old who previously led a company that made money by having poorly paid workers fix AI slop.
Mark Zuckerberg’s gamble on the ‘Metaverse’ has cost Meta more than $77B
The company now plans to cut spending on it and shift focus to AI wearables instead
(via WSJ)
The company now plans to cut spending on it and shift focus to AI wearables instead
(via WSJ)
December 7, 2025 at 3:59 AM
Mind you, the other big Meta bet is making is "Artificial Superintelligence" -- being led by a 28 year old who previously led a company that made money by having poorly paid workers fix AI slop.
December 6, 2025 at 10:24 PM
In a world of newsletters, advertized by how many k words they include, one comes to see the value of editing.
December 6, 2025 at 2:52 PM
In a world of newsletters, advertized by how many k words they include, one comes to see the value of editing.
Reposted by Emily M. Bender
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
twitch.tv
December 6, 2025 at 1:26 AM
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
dair_institute - Twitch
Twitch account for The Distributed AI Research Institute (DAIR).
twitch.tv
December 6, 2025 at 1:26 AM
One problem with all the “AGI” discourse is that “AGI” isn’t actually defined.
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Someone tried recently … and predictably it didn’t go well. Join us for the next Mystery AI Hype Theater 3000 livestream, in which @alexhanna.bsky.social and I dig into the attempt:
Monday, December 8, noon Pacific
Reposted by Emily M. Bender
Bernie Sanders keeps saying the most wrong headed about AI and paling around with Geoffery Hinton. We write more about it here --
buttondown.com/maiht3k/arch...
buttondown.com/maiht3k/arch...
Bernie Sanders Falls for Mathy-Maths and Doomer Gloom
Bernie Sanders published an op-ed in The Guardian this week, warning of the existential dangers posed to civilization by AI.
buttondown.com
December 5, 2025 at 5:21 PM
Bernie Sanders keeps saying the most wrong headed about AI and paling around with Geoffery Hinton. We write more about it here --
buttondown.com/maiht3k/arch...
buttondown.com/maiht3k/arch...
Reposted by Emily M. Bender
🎧 NEW Mystery AI Hype Theater 3000...in which DAIR's own @sophiesong.bsky.social joins @emilymbender.bsky.social and @alexhanna.bsky.social to talk about how some defenders of worker’s rights would rather talk to ChatGPT than, you know, actual workers.
peertube.dair-institute.org/w/xt56PHSUeq...
peertube.dair-institute.org/w/xt56PHSUeq...
December 4, 2025 at 8:17 PM
🎧 NEW Mystery AI Hype Theater 3000...in which DAIR's own @sophiesong.bsky.social joins @emilymbender.bsky.social and @alexhanna.bsky.social to talk about how some defenders of worker’s rights would rather talk to ChatGPT than, you know, actual workers.
peertube.dair-institute.org/w/xt56PHSUeq...
peertube.dair-institute.org/w/xt56PHSUeq...
Reposted by Emily M. Bender
This was such a lovely and refreshing conversation. Ecstatic to see how The AI Con is serving as an organizing tool in conversations around labor, education, health care, and social services.
with @emilymbender.bsky.social
with @emilymbender.bsky.social
"We can and must resist narratives of inevitability through collective labor action and strategic refusal." (The AI Con, p.164)
So thankful to have an opportunity to be in conversation with @alexhanna.bsky.social this evening with @ali-alkhatib.com's book club!
So thankful to have an opportunity to be in conversation with @alexhanna.bsky.social this evening with @ali-alkhatib.com's book club!
December 5, 2025 at 2:14 AM
This was such a lovely and refreshing conversation. Ecstatic to see how The AI Con is serving as an organizing tool in conversations around labor, education, health care, and social services.
with @emilymbender.bsky.social
with @emilymbender.bsky.social
I'd love a browser/general software extension that adds a "No, fuck off!" button to any pop-up suggesting "AI" "help" with something --- and the option to send exactly those words to the company producing the software.
December 4, 2025 at 8:50 PM
I'd love a browser/general software extension that adds a "No, fuck off!" button to any pop-up suggesting "AI" "help" with something --- and the option to send exactly those words to the company producing the software.
Reposted by Emily M. Bender
We are glad to see @sanders.senate.gov pushing back against the tech broligarchy, but we wish he would base his arguments on reality rather than Hinton's fantasies.
www.theguardian.com/commentisfre...
We take apart the report he cites here:
www.theguardian.com/commentisfre...
We take apart the report he cites here:
December 3, 2025 at 2:05 PM
We are glad to see @sanders.senate.gov pushing back against the tech broligarchy, but we wish he would base his arguments on reality rather than Hinton's fantasies.
www.theguardian.com/commentisfre...
We take apart the report he cites here:
www.theguardian.com/commentisfre...
We take apart the report he cites here:
Reposted by Emily M. Bender
Mystery AI Hype Theater 3000 Ep 67:
www.buzzsprout.com/2126417/epis...
… in which @sophiesong.bsky.social joins @alexhanna.bsky.social and me in our dismay at defenders of worker’s rights who’d rather talk to ChatGPT than, you know, actual workers.
Thanks to Ozzy Llinas Goodman for production!
www.buzzsprout.com/2126417/epis...
… in which @sophiesong.bsky.social joins @alexhanna.bsky.social and me in our dismay at defenders of worker’s rights who’d rather talk to ChatGPT than, you know, actual workers.
Thanks to Ozzy Llinas Goodman for production!
December 3, 2025 at 2:01 PM
Mystery AI Hype Theater 3000 Ep 67:
www.buzzsprout.com/2126417/epis...
… in which @sophiesong.bsky.social joins @alexhanna.bsky.social and me in our dismay at defenders of worker’s rights who’d rather talk to ChatGPT than, you know, actual workers.
Thanks to Ozzy Llinas Goodman for production!
www.buzzsprout.com/2126417/epis...
… in which @sophiesong.bsky.social joins @alexhanna.bsky.social and me in our dismay at defenders of worker’s rights who’d rather talk to ChatGPT than, you know, actual workers.
Thanks to Ozzy Llinas Goodman for production!
Reposted by Emily M. Bender
finishing The AI Con by @emilymbender.bsky.social & @alexhanna.bsky.social; i'm so excited we're going to get to chat w/Alex tomorrow. looking forward to hearing if/how the hype landscape has changed, and her observations/thoughts of groups organizing against AI projects like the UM-LANL datacenter.
December 3, 2025 at 11:37 PM
finishing The AI Con by @emilymbender.bsky.social & @alexhanna.bsky.social; i'm so excited we're going to get to chat w/Alex tomorrow. looking forward to hearing if/how the hype landscape has changed, and her observations/thoughts of groups organizing against AI projects like the UM-LANL datacenter.
Reposted by Emily M. Bender
h/t @robertfreundlaw.bsky.social
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content
December 3, 2025 at 7:51 PM
h/t @robertfreundlaw.bsky.social
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content