Being completely open to machine-readability is antithetical to most workplace cultures, and I also suspect it leads to more confident and outspoken people being privileged in decision-making - anything that is too tentative to make the record, or too ambiguous to be understood, will get overlooked
Margaret Mitchell
@mmitchell.bsky.social
Researcher trying to shape AI towards positive outcomes. ML & Ethics +birds. Generally trying to do the right thing. TIME 100 | TED speaker | Senate testimony provider | Navigating public life as a recluse. Former: Google, Microsoft; Current: Hugging Face
This is fascinating. On the work we've done on the impacts of voice transcription, lots of ppl note it can have a chilling effect on workplace relationships (no one wants their start-of-meeting anecdotes logged) and also what they are willing to say "on the record", plus lots of implications for HR
It's worth paying attention not just to what AI can & can't do, but also how companies change the nature of human work in order to make AI more effective. At Anthropic, everyone's work & meetings have to be machine-readable, so that AI agents can see what's happening. www.ft.com/content/9df5...
Thinking LMs are sentient is the flip side of thinking nonhuman animals aren't sentient. We're so attuned to language as a signal for sentience that we can't conceive of unpairing sentience from linguistic competence. So we feel LMs are people, but tool-using crows and octopuses as meat.
Can we please give a HUGE shout out to Stacey Abrams? She worked and fought and organized and created the foundation for the important win by Keisha Lance Bottoms to become the first female and first Black governor in Georgia's history, as well as the first Black female governor in the US.
I appreciate this reporting from @dbeekman.bsky.social in the @seattletimes.com pushing for transparency and clarity in how Seattle's 9-1-1 service is using an automated/speech tech system. >>
Seattle (and other cities) are using AI from a Danish company to triage 911 calls, without serious evaluation. @emilymbender.bsky.social and I give our thoughts. Such technologies are known to struggle with the voices of older and disabled people, i.e., the people most likely to dial 911.
This point from @emilymbender.bsky.social surfaces something that’s been nagging at me: the “stochastic parrot” phrase seems to mostly be used by people trying to say it’s *not* an appropriate term. Why? Why does the term live rent-free in their minds?
I haven't looked deeply at every case, but the overall gestalt is that this remark tends to come from people who are deeply committed to the idea/ideology of "AI" as something all-powerful that will either solve all of our problems or kill us all. >>
This move from Google destroys entire ecosystems. Birds, bugs, plants — all the many things that make our planet habitable — obliterated. www.theguardian.com/environment/...
‘Pure arrogance’: anger in Finland as Google illegally fells 250,000 trees for datacentre
Company ordered to stop work at two sites in country, which has found itself at forefront of tech firms’ latest push
theguardian.com
Listening to a report about Google chopping down a quarter of a million trees in Finland to build a data centre while washing out yoghurt pots for the recycling
Listening to a report about Google chopping down a quarter of a million trees in Finland to build a data centre while washing out yoghurt pots for the recycling
Pleased to announce that my silly TikToks about AI are now officially *major award winning* silly TikToks about AI. :) Thanks to @scicommexcellence.nationalacademies.org for this award but also for helping legitimize/incentivize science communication! www.nationalacademies.org/news/nationa...
"I understand that there have been some experiments showing that LLM review may have similar outcomes to human peer review, but who cares? Either scholarship, research and publishing are human enterprises or they aren’t." www.insidehighered.com/opinion/colu... >>
Generative AI Is Not Our Peer
Automation can make many good things possible, but let’s not pretend it’s something it isn’t.
insidehighered.com
As @biblioracle.bsky.social says, "Either scholarship, research and publishing are human enterprises or they aren’t." We need to talk about why four years in, the conversations that get the most oxygen are about what LLMs can do and not about why we would want them to.
LLMs aren’t “peers” so it makes no sense for them to do peer review. www.insidehighered.com/opinion/colu...
The PHP Foundation, which coordinated fixes for two critical PHP flaws, operates on a shoestring budget with one full-time staff member and 14 contractors (Toby Sterling/Reuters) Main Link | Techmeme Permalink
New @spectrum.ieee.org piece on AI model “welfare”. Well-written, thoughtful article (as always) from @newsbeagle.bsky.social. Super important topic in AI zeitgeist, with input from me, @anilseth.bsky.social and Cameron Berg. Kicker is me talking about bugs ❤️ 🐞 🕷️🪲 spectrum.ieee.org/ai-model-wel...
Extraordinary claims require extraordinary evidence, and I am not the one making the extraordinary claim, so it is not my job to specify what the evidence should be. >>
I found this review of The AI Doc and Ghost in the Machine—two very different films—helpful when trying to separate doomer hype from well-researched documentary filmmaking. The AI Doc falls firmly into doomer hype category, imo. From @deccamuldowney.bsky.social and @emilymbender.bsky.social:
A tale of two "AI" documentaries
By: Decca Muldowney and Emily M. Bender Emily was interviewed for two documentaries which premiered at Sundance this year: The AI Doc: How I Became an...
buttondown.com
RIP Margret Hamilton aka rope mother. Thanks for getting us to the moon. And she did it all without an LLM or even stack overflow. A true legend. news.mit.edu/2026/margare...
Margaret Hamilton, computing pioneer who led software development for the Apollo program, dies at 90
Margaret H. Hamilton, a profoundly influential computer scientist best known for leading the software engineering team at MIT’s Instrumentation Lab during NASA’s Apollo program, died at age 90.
news.mit.edu
This is getting a lot of quotes, so I feel I should correct this a bit (sorry @anildash.com!) “AI” was 100% born toxic! The first AI workshop was born from the field of “automata studies”, and…1/ (Pictured: the folks who convened for the first workshop. Notice anything notable here?)
I can’t emphasize this enough — “AI” wasn’t born corrupt and toxic. For ~75 years, it was mathematicians and scientists doing interesting, thoughtful work and trying to solve problems, making only the usual human mistakes. What “AI” did to machine learning is as tragic as what it did to artists.
I wrote about something that was making me feel nuts: After weeks of discourse around hypothetical existential threats from AI triggered global calls for government intervention, we learned a faulty AI system almost triggered a US-China war. And no one really cared. theintercept.com/2026/09/28/a...
AI Almost Started a U.S.–China War — and No One Seems to Care
The tech world is too busy dreaming up imaginary doomsday scenarios to focus on a very real one that flared up between nuclear powers.
theintercept.com
In the Biden administration, l led the team that developed what was arguably the first White House “AI safety” framework, the AI Bill of Rights. David Robinson, author of this forthright and sobering essay about risk culture at OpenAI and in Silicon Valley was on that team.
New just now in The Atlantic: David Robinson resigned this week. He was among the longest-tenured employees at OpenAI—and oversaw safety reports on 12 frontier launches. He is very worried: “The time for trial and error is over.” You can read his essay here: www.theatlantic.com/technology/2...
A bit from me in @wired.com today wrt our research on AI agents with @evijit.io and Samir Passi. This is a really detailed article, thanks for writing it @thiccreese.bsky.social! www.wired.com/story/metas-...
Meta's Muse Is Better at Surveilling Than Helping Me
The Muse app continues Meta’s trend of opting users into data collection for AI training. It also nudges you to share your bank account, email, and passport information.
wired.com
AI agents make you dumb! 🚨 www.wired.com/story/metas-...
🪶 Captured this great little video of a native Bewick's wren singing to its friends in the forest and it has made my week.
If you read just one thing about the realities of modern AI and agents going "rogue", read this: aiguide.substack.com/p/misleading... Brilliant piece from @melaniemitchell.bsky.social
Misleading Metaphors, Real Risks
What To Fear from AI Agents and How to Reclaim Our Human Agency
aiguide.substack.com
Worth revisiting this prescient paper from @mmitchell.bsky.social and others from @hf.co : arxiv.org/abs/2502.02649
Fully Autonomous AI Agents Should Not be Developed
This paper argues that fully autonomous AI agents should not be developed. In support of this position, we build from prior scientific literature and current product marketing to delineate different A...
arxiv.org
If you are at @deeplearningindaba.bsky.social, remember to check out the Trust AI Workshop!
🚨 Just days away! The 4th TrustAI Workshop @deeplearningindaba.bsky.social , Lagos 🇳🇬 is here! Topic: How trustworthy are LLMs-as-Judges and what does it mean for the African continent? 🎤 Keynote: Dr @chinasa.bsky.social 🗣️ Panel: Tobi Olatunji, Cecilia Mwende, Dr. Blessing Ogbuokiri #AI #LLMs
On August 20, at 1 p.m. ET, the authors will join @mmitchell.bsky.social, chief ethics scientist @hf.co; and Madeleine Claire Elish, head of eval, responsible development, & innovation at Google Deepmind; to discuss their research. Learn more and RSVP! datasociety.net/news-events/...
We @hf.co made an interactive visual of the actual hack from the Hugging Face side: the attack chain across trust boundaries, phase activity, and the commands as they were recorded. Key #transparency . huggingface.co/blog/agent-i... Massive props to Hugo, Adrien, Raphael, Christophe for making this.
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
People not in tech have been interested in what happened with the Open AI agent/Hugging Face hack. So, I "Explain it Like I'm 5"-ed it and made a cartoon in 8 panels. Apologies for some anthro-ing. 🫤 m-mitchell.com/HF-hack-cart... 1/8
If you haven't read this piece on language use in around AI, you definitely should. Great work from @emilymbender.bsky.social , @nannainie.bsky.social and Peter Zukerman, with solutions for what to call things we backoff to just anthropomorphising. firstmonday.org/ojs/index.ph...