Steven Mesquiti
smesquiti.bsky.social
Steven Mesquiti
@smesquiti.bsky.social
PhD Student @Princeton | Interests in AI, Language-based Assessments of Mental Health, & NLP
https://scm1210.github.io/
Pinned
Excited to share my 1st grad school preprint w/ @eriknook.bsky.social
We review how AI is used in mental health research & treatment (from detecting symptoms w/ language to powering therapy chatbots) and highlight ethical/safety issues w/ rapidly bringing AI into this space.
osf.io/preprints/ps...
OSF
osf.io
Reposted by Steven Mesquiti
I've had a missile blow up next to my airplane, been shot at dozens of times by anti-aircraft fire, and launched into orbit — all for my country. I never thought I'd see a President call for my execution.

Trump doesn’t understand the Constitution, and we're all less safe for it.
November 20, 2025 at 8:40 PM
Reposted by Steven Mesquiti
Huge congrats to our lab's summer intern Aryanna Toomer from Howard University who won the undergraduate paper award at the Black in Neuro Conference -- invited to submit a paper on her work to Oxford Open Neuroscience as a freakin sophomore.
November 19, 2025 at 9:14 PM
Reposted by Steven Mesquiti
new paper by Sean Westwood:

With current technology, it is impossible to tell whether survey respondents are real or bots. Among other things, makes it easy for bad actors to manipulate outcomes. No good news here for the future of online-based survey research
November 18, 2025 at 7:16 PM
Reposted by Steven Mesquiti
This post highlights how far behind the OSF is in usability and speed. The ResearchBox UI does seem clean, but why not use Zenodo directly (RB archives to Zenodo anyway.)
November 19, 2025 at 8:14 AM
Reposted by Steven Mesquiti
Hi everyone. A grad in my lab (@rachaelkee.bsky.social) could use your help. Her dog needs surgery. If you can, even a little help goes a long way gofund.me/0ebf78a49
Donate to Help Ninook Walk Again: Emergency TPLO Surgery Needed, organized by Rachael Kee
Hi friends, family, and animal lovers everywhere. My name is Rachael… Rachael Kee needs your support for Help Ninook Walk Again: Emergency TPLO Surgery Needed
gofund.me
November 18, 2025 at 5:52 PM
Reposted by Steven Mesquiti
🚨📣 We’re recruiting a postdoc! 📣🚨
Are you curious about how the environments in which people spend time influence their health? Are you motivated by research that promotes climate action, resilience, and wellbeing?
Read more and apply here: www.asc.upenn.edu/node/5905
Postdoctoral Researcher
www.asc.upenn.edu
November 11, 2025 at 6:28 PM
Reposted by Steven Mesquiti
Global well-being in 2025: A multidimensional analysis of mental, financial, and social health in 92 countries: https://osf.io/3jyda
November 11, 2025 at 9:36 PM
Reposted by Steven Mesquiti
Young voters played a key role in elections across the country last week. It’s fantastic to see record turnout in so many communities AND we still have so much room to grow!

For ideas on how to motivate and support youth to vote check out our new commentary:
osf.io/kyzh9_v1

Key points in 🧵
OSF
osf.io
November 10, 2025 at 6:47 PM
Reposted by Steven Mesquiti
LLMs have been shown to provide different predictions in clinical tasks when patient race is altered. Can SAEs spot this undue reliance on race? 🧵

Work w/ @byron.bsky.social

Link: arxiv.org/abs/2511.00177
November 5, 2025 at 3:20 PM
Reposted by Steven Mesquiti
October 30, 2025 at 10:09 AM
Reposted by Steven Mesquiti
“Each generation has had to learn how to work with its newly acquired cognitive prostheses, whether stylus, scroll, or smartphone,” Kwame Anthony Appiah writes—But what’s different with AI is its speed and the intimacy of the exchange with it. https://theatln.tc/6wgCQscd
October 30, 2025 at 11:20 PM
Reposted by Steven Mesquiti
LLMs are known to perpetuate social biases in clinical tasks. Can we locate and intervene upon LLM activations that encode patient demographics like gender and race? 🧵

Work w/ @arnabsensharma.bsky.social, @silvioamir.bsky.social, @davidbau.bsky.social, @byron.bsky.social

arxiv.org/abs/2502.13319
February 22, 2025 at 4:18 AM
Reposted by Steven Mesquiti
Hey, just want to mumble that this is wildly illegal and inappropriate and would prompt impeachment proceedings if we still had a functional Congress instead of one run by a blind, deaf and mute Renfeld wannabe.
New: The Pentagon confirms to CNN it will funnel $130M from an anonymous Trump “friend” toward military pay.

Asked about the donor’s identity and any foreign or domestic entanglements, the WH referred questions to DoD. DoD then referred questions back to the WH
www.cnn.com/2025/10/24/p...
Pentagon to use $130 million donation from anonymous Trump ‘friend’ to pay military members | CNN Politics
The Trump administration plans to funnel a $130 million donation from an anonymous ally of President Donald Trump toward paying military service members during the government shutdown, the Defense Dep...
www.cnn.com
October 24, 2025 at 7:00 PM
Reposted by Steven Mesquiti
What if we did a single run and declared victory
October 23, 2025 at 2:28 AM
Reposted by Steven Mesquiti
Buried in Meta’s latest round of layoffs were cuts to its team responsible for monitoring risks to user privacy.
Meta Layoffs Included Employees Who Monitored Risks to User Privacy
While the company announced job cuts in artificial intelligence, it also expanded plans to replace privacy and risk auditors with more automated systems.
nyti.ms
October 23, 2025 at 11:40 PM
Reposted by Steven Mesquiti
Great piece on the absurdity of brute force multiverse analyses.

www.pnas.org/doi/10.1073/...
Robustness is better assessed with a few thoughtful models than with billions of regressions | PNAS
Robustness is better assessed with a few thoughtful models than with billions of regressions
www.pnas.org
October 22, 2025 at 5:29 PM
Reposted by Steven Mesquiti
Amazon's cloud computing service provides back-end support to many companies that operate online. When it has problems, so do they. n.pr/47mqC61
Outage at Amazon Web Services disrupts websites across the internet
Amazon's cloud computing service provides back-end support to many companies that operate online. When it has problems, so do they.
n.pr
October 21, 2025 at 1:03 PM
Reposted by Steven Mesquiti
Artificial Intelligence in Mental Health Treatment and Research: https://osf.io/jvf2z
October 11, 2025 at 5:36 AM
Reposted by Steven Mesquiti
LLMs can mirror expert judgment but often rely on word patterns rather than reasoning. A new study introduces epistemia, the illusion of knowledge that occurs when surface plausibility replaces verification. In PNAS: https://ow.ly/ry7S50Xcv9b
October 16, 2025 at 7:00 PM
Reposted by Steven Mesquiti
Check out Steven’s excellent review of an extremely important topic: Views from lots of perspectives are going to be crucial in how we handle this moment in history!
Excited to share my 1st grad school preprint w/ @eriknook.bsky.social
We review how AI is used in mental health research & treatment (from detecting symptoms w/ language to powering therapy chatbots) and highlight ethical/safety issues w/ rapidly bringing AI into this space.
osf.io/preprints/ps...
OSF
osf.io
October 16, 2025 at 2:44 PM
Reposted by Steven Mesquiti
🚨Out now in @cp-trendscognsci.bsky.social 🚨

We explore the use of cognitive theories/models with real-world data for understanding mental health.

We review emerging studies and discuss challenges and opportunities of this approach.

With @yaelniv.bsky.social and @eriknook.bsky.social

Thread ⬇️
September 29, 2025 at 3:05 PM
Excited to share my 1st grad school preprint w/ @eriknook.bsky.social
We review how AI is used in mental health research & treatment (from detecting symptoms w/ language to powering therapy chatbots) and highlight ethical/safety issues w/ rapidly bringing AI into this space.
osf.io/preprints/ps...
OSF
osf.io
October 16, 2025 at 2:30 PM
Reposted by Steven Mesquiti
🚨 New preprint 🚨

Analyzing the academic trajectories of 78,216 psychology researchers, we demonstrate a persistent gender attrition gap, with women psychologists dropping out of academia at consistently higher rates than men psychologists.

Preprint: arxiv.org/pdf/2510.13273
October 16, 2025 at 9:38 AM
Reposted by Steven Mesquiti
50% of American adults are more concerned than excited about the increased use of AI in daily life. Just 10% are more excited than concerned.
🧪
October 15, 2025 at 4:21 PM
Reposted by Steven Mesquiti
Every targeted institution (my own very much included) should cut and paste this letter onto their letterhead.
A master class from MIT in responding to authoritarian overreach:

Your “premise … is inconsistent with our core belief that scientific funding should be based on scientific merit alone.
… America’s leadership in science & innovation depends on independent thinking & open competition for excellence.
October 10, 2025 at 3:24 PM