Emily M. Bender
LightNews LightNews
banner
emilymbender.bsky.social
Emily M. Bender
@emilymbender.bsky.social
37K followers 580 following 2.6K posts
Book: https://thecon.ai

Web: https://faculty.washington.edu/ebender
Posts Replies Media Videos
Reposted by Emily M. Bender
alexhanna.bsky.social
AI is not a “team of geniuses”. Sam Altman says that (unnamed) scientists feel as though they are marshalling the knowledge of a whole “AI” team. But science doesn’t work like that.

(with @emilymbender.bsky.social)
January 3, 2026 at 11:06 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
If Rosenberg had actually cited the people he's arguing against, and made it possible to click through and see their words, I'm pretty sure what you'd find is not denialism, but resistance.

>>
January 3, 2026 at 5:47 AM
Reposted by Emily M. Bender
emilymbender.bsky.social
Not sure why this dross, dated Dec 1, seems to be circulating now (and why it didn't cross my feed a month ago), but wow what a terrible essay.

A few comments, in a short 🧵>>

bigthink.com/the-present/...
The rise of AI denialism
Computer scientist Louis Rosenberg argues that dismissing AI as a “bubble” or mere “slop” ignores the tectonic technological shift now unfolding.
bigthink.com
January 3, 2026 at 5:25 AM
emilymbender.bsky.social
Not sure why this dross, dated Dec 1, seems to be circulating now (and why it didn't cross my feed a month ago), but wow what a terrible essay.

A few comments, in a short 🧵>>

bigthink.com/the-present/...
The rise of AI denialism
Computer scientist Louis Rosenberg argues that dismissing AI as a “bubble” or mere “slop” ignores the tectonic technological shift now unfolding.
bigthink.com
January 3, 2026 at 5:25 AM
Reposted by Emily M. Bender
emilymbender.bsky.social
This zine was created on a quiet, meditative afternoon in November in Seattle, at a workshop led by @dairinstitute.bsky.social 's Dylan Baker -- a lovely experience and a beautiful result!

zines.dair-institute.org/possible-fut...
The DAIR Zine Library
A collection of zines from the DAIR Institute
zines.dair-institute.org
January 2, 2026 at 5:10 AM
emilymbender.bsky.social
This zine was created on a quiet, meditative afternoon in November in Seattle, at a workshop led by @dairinstitute.bsky.social 's Dylan Baker -- a lovely experience and a beautiful result!

zines.dair-institute.org/possible-fut...
The DAIR Zine Library
A collection of zines from the DAIR Institute
zines.dair-institute.org
January 2, 2026 at 5:10 AM
Reposted by Emily M. Bender
techpolicypress.bsky.social
Most-listened Tech Policy Press podcasts of 2025: In June, Justin Hendrix spoke to Emily M. Bender and Alex Hanna about their new book, The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want.
Taking on the AI Con | TechPolicy.Press
Emily M. Bender and Alex Hanna are the authors of The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want, just out from Harper Collins.
buff.ly
January 1, 2026 at 4:54 PM
Reposted by Emily M. Bender
gretchenmcc.bsky.social
If you know a high school student interested in language and/or puzzles, tell 'em about the linguistics olympiads, there are lots of national versions in many countries plus an international competition!
linguisticdiscovery.com Linguistic Discovery @linguisticdiscovery.com · 2d
The International Linguistics Olympiad (IOL) is an annual international competition that brings together secondary school students and experts from various fields of linguistics.
January 1, 2026 at 8:39 PM
Reposted by Emily M. Bender
vocalfriespod.bsky.social
1. The AI Con, where we talk with @emilymbender.bsky.social and @alexhanna.bsky.social about their book, The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want

www.buzzsprout.com/1930383/epis...
The AI Con - The Vocal Fries
Carrie and Megan talk with Drs Emily M. Bender, Professor of Linguistics at the University of Washington, and Alex Hanna, Director of Research at the Distributed AI Research Institute (DAIR), about th...
www.buzzsprout.com
January 1, 2026 at 5:31 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
MAIHT3k, Ep 69:

www.buzzsprout.com/2126417/epis...

I thought it would take at least some effort to tear this paper-shaped object apart. Instead, it was so flimsy it disintegrated under the force of @alexhanna.bsky.social & I simply … reading it.

Thanks to Ozzy Llinas Goodman for production!
December 31, 2025 at 2:00 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
Also available as video on PeerTube:
peertube.dair-institute.org/w/cNdmpvRHXW...
December 31, 2025 at 2:01 PM
emilymbender.bsky.social
明けましておめでとうございます
Photo of Senso-ji, the temple at Asakusa, Tokyo, at night with a crowd in front of it
December 31, 2025 at 3:25 PM
emilymbender.bsky.social
Also available as video on PeerTube:
peertube.dair-institute.org/w/cNdmpvRHXW...
December 31, 2025 at 2:01 PM
emilymbender.bsky.social
MAIHT3k, Ep 69:

www.buzzsprout.com/2126417/epis...

I thought it would take at least some effort to tear this paper-shaped object apart. Instead, it was so flimsy it disintegrated under the force of @alexhanna.bsky.social & I simply … reading it.

Thanks to Ozzy Llinas Goodman for production!
December 31, 2025 at 2:00 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
If someone wants to be dismissive of those values, that is a choice they can make, and I think by taking the initiative to frame the debate and making clear what is at stake, we can show that choice for the dangerous, anti-social one that it is.

/fin
December 30, 2025 at 12:10 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
I enjoyed doing this interview for El Mundo, especially the last question (English translation of that in this short thread):

www.elmundo.es/papel/histor...

🧵>>
Emily M. Bender: "Las empresas pueden utilizar la excusa de la IA para perjudicar a los trabajadores de muchas maneras"
Emily M. Bender es una de las voces más críticas y lúcidas en el debate sobre la inteligencia artificial actual. Profesora de Lingüística en la Universidad de Washington,...
www.elmundo.es
December 30, 2025 at 12:08 PM
Reposted by Emily M. Bender
emilymbender.bsky.social
My reply:

Sometimes, we will be described as "anti-progress" and the first step is to accept that that might happen, but that refusal is important nonetheless. But being described as "anti-progress" doesn't necessarily mean being "dismissed".

>>
December 30, 2025 at 12:09 PM
Reposted by Emily M. Bender
irisvanrooij.bsky.social
“Those adherents to the myth of AGI believe that the only and best thing humans can do right now is work on a superintelligence, which will bring about a new age of abundance.” — @alexhanna.bsky.social and @emilymbender.bsky.social

www.techpolicy.press/the-myth-of-...
The Myth of AGI | TechPolicy.Press
Alex Hanna and Emily M. Bender write that claims of "Artificial General Intelligence" are a cover for abandoning the current social contract.
www.techpolicy.press
December 30, 2025 at 10:09 PM
emilymbender.bsky.social
I enjoyed doing this interview for El Mundo, especially the last question (English translation of that in this short thread):

www.elmundo.es/papel/histor...

🧵>>
Emily M. Bender: "Las empresas pueden utilizar la excusa de la IA para perjudicar a los trabajadores de muchas maneras"
Emily M. Bender es una de las voces más críticas y lúcidas en el debate sobre la inteligencia artificial actual. Profesora de Lingüística en la Universidad de Washington,...
www.elmundo.es
December 30, 2025 at 12:08 PM
Reposted by Emily M. Bender
alexhanna.bsky.social
Honored that our article, "The Myth of AGI", was one of Tech Policy Press's Top 30 read pieces of the year.

In some great company here, especially with @eryk.bsky.social's "Anatomy of an AI Coup".

www.techpolicy.press/top-30-most-...

(with @emilymbender.bsky.social)
Top 30 Most Read Pieces on Tech Policy Press in 2025 | TechPolicy.Press
In 2025, Tech Policy Press published over 1,100 posts, including articles, analyses, perspectives, transcripts, trackers, podcasts, and more.
www.techpolicy.press
December 29, 2025 at 10:42 PM
Reposted by Emily M. Bender
alexhanna.bsky.social
Don't forget that Ilya cited this as THE
usecase he was most excited about.
A tweet from Ilya Sustkever. It reads, "In the future, once the robustness of our models will exceed some threshold, we will have *wildly effective* and dirt cheap AI therapy. Will lead to a radical improvement in people's experience of life. One of the applications I'm most early awaiting."
December 29, 2025 at 2:07 AM
Reposted by Emily M. Bender
alexhanna.bsky.social
From @theinformation.com newsletter -- "Applied AI: Microsoft Offers AI Copilot Customers Money to Train Employees to Use It"

"To bridge the gap, Microsoft is now offering some large customers money for employee training when they purchase a large number of 365 Copilot seats."
Screenshot from The Information newsletter, titled "Applied AI". By Aaron Holmes. Microsoft CEO Satya Nadella and his deputies have warned that plenty of employees at large enterprises that pay for Microsoft 365 Copilot haven’t been using the AI features very much, as we reported recently. The most obvious way to change their ways is to make the software, an artificial intelligence add-on to Office apps like Word and Excel, more reliable and useful, which is what Nadella is personally focused on. The other is to teach workers how to use the features, as many of them just aren’t good at it, IT executives say. To bridge the gap, Microsoft is now offering some large customers money for employee training when they purchase a large number of 365 Copilot seats. That’s what Microsoft did for public sector employees across North Carolina, including at the City of Raleigh, which recently agreed to purchase around 200 seats of 365 Copilot, according to Raleigh Chief Information Officer Mark Wittenburg.
December 29, 2025 at 9:00 PM
emilymbender.bsky.social
Wait, so per @tdietterich.bsky.social authors shouldn't use LLMs (ahem, without checking) to write papers, but the first pass of review can be fully automated with 'AI' (meaning, I'd guess, LLMs)?

I mean obviously desk reject LLM slop, but also don't pawn off peer review to LLM slop.
Comment by Tom Diettrich on a linkedin post reading: "You can't "test-in quality" in engineering; you can't "review-in quality" in research. We need incentives for people to do better research. Our system today assumes that 75% of submitted papers are low quality, and it is probably right (I'll bet it is higher). If this were a manufacturing organization, an 75% defect rate would result in bankruptcy. Imagine a world in which you could have an AI system check the correctness/quality of your paper. If your paper passed that bar, then it could be published (say, on arXiv). Subsequent human review could assess its importance to the field. In such a system, authors would be incentivized to satisfy the AI system. This will lead to searching for exploits in the AI system. A possible solution is to select the AI evaluator at random from a large pool and limit the number of permitted submissions. I imagine our colleagues in mechanism design can improve on this idea." Original: https://www.linkedin.com/feed/update/urn:li:activity:7381685800549257216/?commentUrn=urn%3Ali%3Acomment%3A(activity%3A7381685800549257216%2C7382628060044599296)&dashCommentUrn=urn%3Ali%3Afsd_comment%3A(7382628060044599296%2Curn%3Ali%3Aactivity%3A7381685800549257216) Start of a bsky thread by Tom Diettrich reading: I'd like to propose the following norm for peer review of papers. If a paper shows clear signs of LLM-generated errors that were not detected by the author, the paper should be immediately rejected. My reasoning: 1/ #ResearchIntegrity If the authors didn't notice the errors, they probably didn't notice other errors in the paper; therefore we can't trust anything in the paper. This would act as a powerful incentive for authors to check LLM-generated material very carefully 2/ Original: https://bsky.app/profile/tdietterich.bsky.social/post/3mazoofislc2m
December 28, 2025 at 10:48 PM
Reposted by Emily M. Bender
bookjockeyalex.bsky.social
"Academics literally cannot make genAI go away" we also can't make underage drinking go away and we're not advocating for installing bar carts in every classroom.
December 27, 2025 at 11:46 PM
emilymbender.bsky.social
Under what circumstances is it acceptable to just not reply to requests to review from journals I've never heard of that are furthermore outside my field?

#AcademicSky
December 28, 2025 at 3:21 AM