M.J. Crockett
@mjcrockett.bsky.social
Professor of Psychology & Human Values at Princeton | Cognitive scientist curious about technology, narratives, & epistemic (in)justice | They/She 🏳️🌈
www.crockettlab.org
www.crockettlab.org
Pinned
AI Surrogates and illusions of generalizability in cognitive science
Recent advances in artificial intelligence (AI) have generated enthusiasm for using AI simulations of human research participants to generate new know…
www.sciencedirect.com
Can AI simulations of human research participants advance cognitive science? In @cp-trendscognsci.bsky.social, @lmesseri.bsky.social & I analyze this vision. We show how “AI Surrogates” entrench practices that limit the generalizability of cognitive science while aspiring to do the opposite. 1/
Reposted by M.J. Crockett
hi, can you do me a favor?
take a picture of your pet and then take a picture again while you sing to them
thank you, this is for science
here is my contribution
take a picture of your pet and then take a picture again while you sing to them
thank you, this is for science
here is my contribution
November 10, 2025 at 11:25 PM
hi, can you do me a favor?
take a picture of your pet and then take a picture again while you sing to them
thank you, this is for science
here is my contribution
take a picture of your pet and then take a picture again while you sing to them
thank you, this is for science
here is my contribution
Brilliant
New art video just dropped! The 7 Trillion Dollar Seance 🔮 This is meant to be performed live, recited in front of the video at conferences, museums, public events. It got a standing ovation at Tech Together in NYC. Your phone works but best on your largest screen.
vimeo.com/1116289621
🧵
vimeo.com/1116289621
🧵
The 7 Trillion Dollar Seance
Şerife Wong The 7 Trillion Dollar Seance 2025 Original score by Jeromey Cooks (rikuwru) This video is part of Icarus Salon. It is intended to be viewed in…
vimeo.com
November 11, 2025 at 2:27 AM
Brilliant
Reposted by M.J. Crockett
Hit the breaking point this morning with Michigan after hearing about their planned data center in Ypsi. Put all my in-progress posts on the back burner to write this (which I have also sent to UMGiving).
substack.com/@kconrad/not...
substack.com/@kconrad/not...
A letter to my alma mater
on their embrace of AI
substack.com
November 10, 2025 at 8:15 PM
Hit the breaking point this morning with Michigan after hearing about their planned data center in Ypsi. Put all my in-progress posts on the back burner to write this (which I have also sent to UMGiving).
substack.com/@kconrad/not...
substack.com/@kconrad/not...
Reposted by M.J. Crockett
Just read a talk announcement about "AI" and K-12 education that looked like it maybe, maybe embedded a critical perspective but was still dressed up in the language of AI hype, presupposing both job market & education "reshaped by AI".
🧵>>
🧵>>
November 10, 2025 at 8:32 PM
Just read a talk announcement about "AI" and K-12 education that looked like it maybe, maybe embedded a critical perspective but was still dressed up in the language of AI hype, presupposing both job market & education "reshaped by AI".
🧵>>
🧵>>
This NeurIPS workshop claims that LLMs "provide an important foundation for exploring human cognition, emotion, and social interaction"
This is flawed logic, as @lmesseri.bsky.social and I argue here:
www.sciencedirect.com/science/arti...
This is flawed logic, as @lmesseri.bsky.social and I argue here:
www.sciencedirect.com/science/arti...
November 9, 2025 at 3:23 PM
This NeurIPS workshop claims that LLMs "provide an important foundation for exploring human cognition, emotion, and social interaction"
This is flawed logic, as @lmesseri.bsky.social and I argue here:
www.sciencedirect.com/science/arti...
This is flawed logic, as @lmesseri.bsky.social and I argue here:
www.sciencedirect.com/science/arti...
Important thread 👇
I appreciate the work of these authors to show that this problem not only is still here but has grown:
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
AI's capabilities may be exaggerated by flawed tests, according to new study
A study from the Oxford Internet Institute analyzed 445 tests used to evaluate AI models.
www.nbcnews.com
November 9, 2025 at 3:20 PM
Important thread 👇
Reposted by M.J. Crockett
"This long-running war on knowledge and expertise has sown the ground for the narratives widely used by AI companies and the CEOs adopting it. Human labor, inquiry, creativity, and expertise is spurned in the name of “efficiency.” With AI, there is no need for human expertise"
AI Is Supercharging the War on Libraries, Education, and Human Knowledge
"Fascism and AI, whether or not they have the same goals, they sure are working to accelerate one another."
www.404media.co
November 7, 2025 at 4:20 PM
"This long-running war on knowledge and expertise has sown the ground for the narratives widely used by AI companies and the CEOs adopting it. Human labor, inquiry, creativity, and expertise is spurned in the name of “efficiency.” With AI, there is no need for human expertise"
Reposted by M.J. Crockett
New — While some states have received SNAP benefits for November and USDA is supposedly pushing through all federal payments soon, that doesn’t erase the psychological terror of the last few weeks.
I spoke with 20+ SNAP recipients in 16 states about preparing for a possible future with no food:
I spoke with 20+ SNAP recipients in 16 states about preparing for a possible future with no food:
The voices of SNAP
Recipients have become political pawns. They explained, in their own words, what Trump withholding funds has been like.
www.thehandbasket.co
November 7, 2025 at 10:55 PM
New — While some states have received SNAP benefits for November and USDA is supposedly pushing through all federal payments soon, that doesn’t erase the psychological terror of the last few weeks.
I spoke with 20+ SNAP recipients in 16 states about preparing for a possible future with no food:
I spoke with 20+ SNAP recipients in 16 states about preparing for a possible future with no food:
Reposted by M.J. Crockett
Panera’s moderately caffeinated lemonade was loosely associated with 2 deaths before it was taken off market.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
I fully believe in the corporate death penalty and believe we would be a better world if OpenAI lost its corporate charter and was forcibly dissolved.
www.cnn.com/2025/11/06/u...
www.cnn.com/2025/11/06/u...
ChatGPT encouraged college graduate to commit suicide, family claims in lawsuit against OpenAI | CNN
A 23-year-old man killed himself in Texas after ChatGPT ‘goaded’ him to commit suicide, his family says in a lawsuit.
www.cnn.com
November 7, 2025 at 10:56 PM
Panera’s moderately caffeinated lemonade was loosely associated with 2 deaths before it was taken off market.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
Reposted by M.J. Crockett
Excited to welcome @lmesseri.bsky.social, @cameronbuckner.bsky.social and @carlbergstrom.com to campus a week from today for the “AI and the Nature of Science: Concepts and Controversies” event. 14.11.25, 1400 - 1800.
AI and the Nature of Science
Bringing together cutting-edge perspectives and helping set an agenda for future research and application
buff.ly
November 7, 2025 at 2:45 PM
Excited to welcome @lmesseri.bsky.social, @cameronbuckner.bsky.social and @carlbergstrom.com to campus a week from today for the “AI and the Nature of Science: Concepts and Controversies” event. 14.11.25, 1400 - 1800.
Reposted by M.J. Crockett
congrats to everyone for thinking this through before signing very big contracts to put it on every student’s device at the school or university you run
OVER A MILLION USERS
DISCUSS SUICIDE WITH CHATGPT
EVERY *WEEK*
what the fuck are we DOING here
DISCUSS SUICIDE WITH CHATGPT
EVERY *WEEK*
what the fuck are we DOING here
Panera’s moderately caffeinated lemonade was loosely associated with 2 deaths before it was taken off market.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
This article alone has 4 examples of ChatGPT encouraging young people to commit suicide, and OpenAI’s own public stats estimate over a million users discuss suicide with ChatGPT each week.
November 7, 2025 at 11:24 PM
congrats to everyone for thinking this through before signing very big contracts to put it on every student’s device at the school or university you run
Pop-up AI defacing your article critical of AI captures a lot of what it feels like to work in this space.
@lmesseri.bsky.social
@lmesseri.bsky.social
November 7, 2025 at 3:47 PM
Pop-up AI defacing your article critical of AI captures a lot of what it feels like to work in this space.
@lmesseri.bsky.social
@lmesseri.bsky.social
Important work revealing limitations of using LLMs as human surrogates.
Note, however, that even if LLMs' textual outputs were perfectly human-like, they would still be poor models of human cognition, as @lmesseri.bsky.social argue here:
www.sciencedirect.com/science/arti...
Note, however, that even if LLMs' textual outputs were perfectly human-like, they would still be poor models of human cognition, as @lmesseri.bsky.social argue here:
www.sciencedirect.com/science/arti...
November 7, 2025 at 3:39 PM
Important work revealing limitations of using LLMs as human surrogates.
Note, however, that even if LLMs' textual outputs were perfectly human-like, they would still be poor models of human cognition, as @lmesseri.bsky.social argue here:
www.sciencedirect.com/science/arti...
Note, however, that even if LLMs' textual outputs were perfectly human-like, they would still be poor models of human cognition, as @lmesseri.bsky.social argue here:
www.sciencedirect.com/science/arti...
Reposted by M.J. Crockett
Another person who started down this path from the most “innocuous” advertised uses of ChatGPT: homework help. I have been struck by the cases of people getting lost in delusions after using this product for work or school: ie the use cases we are supposed to encourage as faculty.
November 7, 2025 at 12:50 PM
Another person who started down this path from the most “innocuous” advertised uses of ChatGPT: homework help. I have been struck by the cases of people getting lost in delusions after using this product for work or school: ie the use cases we are supposed to encourage as faculty.
Reposted by M.J. Crockett
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...
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
November 7, 2025 at 4:14 AM
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...
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...
Reposted by M.J. Crockett
1.This is not the only known example we know of ChatGPT cheering on a vulnerable young person to die by suicide, this time it is a recent Texas A&M grad
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
November 7, 2025 at 1:50 AM
1.This is not the only known example we know of ChatGPT cheering on a vulnerable young person to die by suicide, this time it is a recent Texas A&M grad
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
Reposted by M.J. Crockett
There are no words for how evil this is
November 7, 2025 at 1:36 AM
There are no words for how evil this is
Reposted by M.J. Crockett
November 6, 2025 at 1:09 AM
Reposted by M.J. Crockett
ML based facial recognition is not reliable in the wild; like many “AI” tools sold to law enforcement, it is thinly-veiled pseudoscience and will harm people by lending law enforcement both a felt sense of epistemic authority and legal impunity.
Could errors from an AI surveillance tool cause ICE to grab and detain you? The agency’s use of facial recognition — and alarming new details on its reckless procedures for doing so — make that sci-fi-sounding nightmare an all too real risk, writes CDT surveillance expert Jake Laperruque.
ICE’s Reckless Reliance on Facial Recognition Puts Us All In Danger | TechPolicy.Press
Jake Laperruque raises the alarm on ICE’s reckless use of facial recognition and its risks for wrongful detainment.
www.techpolicy.press
November 5, 2025 at 8:51 PM
ML based facial recognition is not reliable in the wild; like many “AI” tools sold to law enforcement, it is thinly-veiled pseudoscience and will harm people by lending law enforcement both a felt sense of epistemic authority and legal impunity.
Reposted by M.J. Crockett
Two great scholars doing work on moral (@drmeltemyucel.bsky.social) & personality (@tedmond.bsky.social) development & change are recruiting PhD students this year @msupsychology.bsky.social!
Your grad office could be across the hall from my office!*
*not necessarily a selling point, just a fact
Your grad office could be across the hall from my office!*
*not necessarily a selling point, just a fact
Come work with us @msupsychology.bsky.social!
I promise, we don’t match *all* the time @tedmond.bsky.social
I promise, we don’t match *all* the time @tedmond.bsky.social
November 5, 2025 at 2:22 PM
Two great scholars doing work on moral (@drmeltemyucel.bsky.social) & personality (@tedmond.bsky.social) development & change are recruiting PhD students this year @msupsychology.bsky.social!
Your grad office could be across the hall from my office!*
*not necessarily a selling point, just a fact
Your grad office could be across the hall from my office!*
*not necessarily a selling point, just a fact
Reposted by M.J. Crockett
To understand what was at stake here: in VA, the AG gets to appoint university counsel. From Jim Ryan’s forced resignation to the (mercifully) reject “Compact” to the odious and bizarre “settlement,” UVA has been hamstrung by counsel chosen by the state’s GOP administration. That’s about to change.
November 5, 2025 at 1:52 AM
To understand what was at stake here: in VA, the AG gets to appoint university counsel. From Jim Ryan’s forced resignation to the (mercifully) reject “Compact” to the odious and bizarre “settlement,” UVA has been hamstrung by counsel chosen by the state’s GOP administration. That’s about to change.
Reposted by M.J. Crockett
The lesson of history is that tyrants cannot be appeased.
Appeasement only feeds their insatiable hunger for power.
Tyrants prey on people who feel unsafe and vulnerable.
But without mass submission, a tyrant is powerless.
Remember this.
Appeasement only feeds their insatiable hunger for power.
Tyrants prey on people who feel unsafe and vulnerable.
But without mass submission, a tyrant is powerless.
Remember this.
November 5, 2025 at 5:28 AM
The lesson of history is that tyrants cannot be appeased.
Appeasement only feeds their insatiable hunger for power.
Tyrants prey on people who feel unsafe and vulnerable.
But without mass submission, a tyrant is powerless.
Remember this.
Appeasement only feeds their insatiable hunger for power.
Tyrants prey on people who feel unsafe and vulnerable.
But without mass submission, a tyrant is powerless.
Remember this.
Reposted by M.J. Crockett
For years, but especially for the past year, Democrats have been in thrall to the idea that they need to chase an imagined median voter, and adopt that voter’s positions. Mamdani’s victory shows another way: hold principles, and persuade people to share them.
I wrote about Zohran’s victory. www.theguardian.com/commentisfre...
Mamdani’s victory is a rebuke to the failed strategies of the Democratic party | Moira Donegan
The Democratic party appears listless and unprincipled, unwilling to fight because they do not believe in anything. Zohran Mamdani is the opposite of this
www.theguardian.com
November 5, 2025 at 6:13 AM
For years, but especially for the past year, Democrats have been in thrall to the idea that they need to chase an imagined median voter, and adopt that voter’s positions. Mamdani’s victory shows another way: hold principles, and persuade people to share them.
Reposted by M.J. Crockett
"[Chicago mom] Andrea Soria... overheard her 6-year-old whisper to her dolls: 'We have to be good or ICE will get us.'"
apnews.com/article/chic...
apnews.com/article/chic...
Chicago's children are getting caught in the chaos of immigration crackdowns
In Chicago, children and families have been traumatized by an immigration enforcement action involving tear gas.
apnews.com
November 3, 2025 at 5:42 PM
"[Chicago mom] Andrea Soria... overheard her 6-year-old whisper to her dolls: 'We have to be good or ICE will get us.'"
apnews.com/article/chic...
apnews.com/article/chic...