Margaret Mitchell
banner
mmitchell.bsky.social
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 .. more

Margaret Mitchell is a computer scientist who works on algorithmic bias and fairness in machine learning. She is most well known for her work on automatically removing undesired biases concerning demographic groups from machine learning models, as well as more transparent reporting of their intended use. .. more

Computer science 94%
Psychology 5%

Also love "behavior hijacking".

That's a good one! Maybe the external-facing version of this is Gen-Juicing.

Oh that's great. I actually find that hard to scroll past, especially on my phone, since it tends to take up a lot of space, and THEN it's similar questions that upon dropdown click ALSO have AI-generated responses, and THEN it's finally links.
Have you proactively disabled some AI things?

Yes 100%!

Reposted by Margaret Mitchell

If you have been framing your work as involving/in relation to "AI", what do you mean by "AI"? How would you describe your work without using that phrase?

>>

Haha. As a vegetarian, I particularly like that one.

There's this form of behavioral engineering/coercion going on to use AI when you're not even trying to. Ex: Enterprise Google Slides replaced "Upload an Image", with "Generate an Image" (right?), requiring extra work +adaptation to sidestep the AI-as-default push. It drives me bonkers.

Love the phrases "Cognitive Cost" and "Executive Function Theft" to pinpoint how exhausting it is to be constantly bombarded with apps telling you to use AI. I've been calling it "Corporate pressure" and "Force feeding". It's quite a lot right now. Seems a bit desperate tbh.
The #ExecutiveFunctionTheft of having to opt out.
Feeling annoyed at the cognitive cost of having to dismiss all the offers of "AI" assitance every time I use Acrobat to provide feedback on documents my students wrote. (NO I do NOT want an "AI" summary of this. In what world???)

Decided to check settings:

Reposted by Margaret Mitchell

The #ExecutiveFunctionTheft of having to opt out.
Feeling annoyed at the cognitive cost of having to dismiss all the offers of "AI" assitance every time I use Acrobat to provide feedback on documents my students wrote. (NO I do NOT want an "AI" summary of this. In what world???)

Decided to check settings:

Reposted by Margaret Mitchell

In this 2025 study, more than half of LLM responses cited fake sources or links.

www.cjr.org/tow_center/w...
AI Search Has A Citation Problem
We Compared Eight AI Search Engines. They’re All Bad at Citing News.
www.cjr.org

🪶 Golden-crowned Kinglets at the water spout.

Was it “Q”? I never trust that one.

Reposted by Margaret Mitchell

Vince Gilligan on AI:

“AI is the world’s most expensive and energy-intensive plagiarism machine. I think there’s a very high possibility that this is all a bunch of horseshit. It’s basically a bunch of centibillionaires whose greatest life goal is to become the world’s first trillionaires.”
Vince Gilligan and Rhea Seehorn on What ‘Pluribus’ Is Really About, Why Hollywood ‘Needs More Heroes’ and How Silicon Valley Has ‘F—ed Up the World’
Vince Gilligan and Rhea Seehorn unpack the first two episodes of 'Pluribus' and slam AI: 'Silicon Valley' has 'f---ed up the world.'
variety.com
Holy shit. Noam Shazeer, one of the original authors on the "Attention is All You Need" paper and Character.AI founder, came out as major transphobe. Like Trumpian levels of "this is child mutilation" of transphobia.

(via The Information)

Reposted by Joanna Bryson

This.
When fines are treated by companies as a cost of doing business, the regulatory model itself has failed. #regulatorycapture
When fines are treated by companies as a cost of doing business, the regulatory model itself has failed. #regulatorycapture
When deception’s more profitable than honesty, users lose.

Internal documents suggest that Meta earns $3.5B(!) every 6 mo. from scam ads, revealing a deeper systemic problem: the economic incentive to tolerate “higher legal risk” content still outweighs any penalty. www.reuters.com/investigatio...
Meta is earning a fortune on a deluge of fraudulent ads, documents show
Meta projected 10% of its 2024 revenue would come from ads for scams and banned goods, and it internally estimates that its platforms show users 15 billion scam ads a day, company documents show.
www.reuters.com

I thought this headline was from some small newspaper and went to tag @nytpitchbot.bsky.social to share smthn even more ridiculous than the usual satirical NY Times pitches, but…Nope, this is actually from @nytimes.com
the question nobody with a soul is asking, answered by people nobody wants to hear from

💔
the question nobody with a soul is asking, answered by people nobody wants to hear from
This really hasn’t gotten enough attention this week.

@wired.com #ICE
www.wired.com/story/fbi-wa...

Reposted by Margaret Mitchell

"'Luddite' isn’t an insult—it’s a warning and a word of caution. One we must return to, repeatedly, if we’re to resist the constant drumbeat of consumerism dressed up as science and progress."

Erik J. Larson reviews "The AI Con": https://lareviewofbooks.org/article/the-return-of-the-luddites/

This for me is the fundamental reason why I came to Bsky — posting on X helps retain people on X.
I’ve learned for myself and other colleagues in AI tho that part of the tradeoff w Bsky is abuse: how much abuse is ok for staying on the platform? Are death threats ok? (There isn’t one right answer.)

Yah, sorry. My sense from the comments I saw last night/this morning is that there’s nuance not well-suited for social media convo, this included.
I’m being called a c*** and stuff for things I haven’t said or done, so gotta kinda mute this stuff now—thanks for your thoughtful comments!

100%, no disagreement. I think there are a few issues getting conflated here. Block lists are useful and people should absolutely be able to *not* see people they don't want to see on social media. I disagree a bit with Will on his point -- but he's not speaking for me.

Btw, I think it's important to call it "X" -- not Twitter -- because the "Twitter" moniker gives it an air of legitimacy it doesn't have anymore.

Heya, you tagged me so I feel like I should clarify I haven't left Bsky, but I did start reading X again. And I haven't advocated for people to listen to me who don't want to listen to me. I've seen a few of these messages go around but there seems like there are a lot of issues getting mixed up. 1/

In the US, today is the day: Vote!