Today I joined more than 100 experts in signing a letter calling for 5 minimum requirements to make embedded evaluations credible: aievaluatorforum.org/initiatives/...

Minimum Conditions for Embedding Evaluators

An open letter calling on all frontier AI companies to embed independent evaluators — signed by 100+ researchers, practitioners, and organization leaders.

aievaluatorforum.org

🌟 NEW PAPER PREPRINT 🌟 "I Felt Very Seen, But Still Very Alone": For 6 months, we followed people who use LLMs (e.g., ChatGPT, Claude, etc.) for mental/emotional support to track how it changed 🧵 Full paper here: arxiv.org/abs/2609.123... @merylye.bsky.social, @livgar.bsky.social, Ranjit Singh

"I Felt Very Seen, But Still Very Alone": Longitudinal Trajectories of General-Purpose LLM Use for Socioemotional Support

People increasingly use general-purpose chatbots such as ChatGPT, Claude, and Gemini for mental health and emotional support. We report a multi-stage longitudinal qualitative study of 18 U.S. adults, ...

arxiv.org

arXiv cs.HC Human-Computer Interaction@cshc-bot.bsky.social · last mo.

Meryl Ye, Briana Vecchione, Livia Garofalo, Ranjit Singh: "I Felt Very Seen, But Still Very Alone": Longitudinal Trajectories of General-Purpose LLM Use for Socioemotional Support https://arxiv.org/abs/2609.12314 https://arxiv.org/pdf/2609.12314 https://arxiv.org/html/2609.12314

New from 404 Media: humans are reading ChatGPT conversations. OpenAI has hired an army of contractors who read real ChatGPT users' chats. I've seen internal docs, the review system, and real user prompts. Sometimes they contain very personal and sensitive information www.404media.co/inside-proje...

Inside ‘Project Lily’: The Humans Reading Your ChatGPT Chats

Humans are reading ChatGPT users’ prompts to improve OpenAI’s models, and those chats can include sensitive, personal information, according to leaked internal documents and real prompts seen by 404 M...

404media.co

Join us on June 4 @ 12:30 pm ET: In an event produced with our partners @pulitzercenter.org, D&S researchers @livgar.bsky.social & @briana-v.bsky.social and Pulitzer Center Fellow @paticlarke.bsky.social will discuss chatbots and mental health! pulitzercenter.org/event/invest...

Investigating the Impact of Chatbots on Mental Wellbeing

Image How is engagement with chatbots impacting the mental wellbeing of users worldwide? What needs to be better understood about the ways that chatbots are designed and how interactions with chatbots...

pulitzercenter.org

🌟 NEW PREPRINT 🌟 After six months of following people using chatbots for socioemotional health & 1 year of participatory engagements, we're proud to release a longitudinal qualitative study on how people use general-purpose LLMs and what it demands of AI governance: arxiv.org/abs/2605.23787

Engagement-Optimized Care: When LLMs become Mental Health Infrastructure

General-purpose LLMs are increasingly functioning as mental health infrastructure due to gaps in care left by provider shortages, inadequate insurance coverage, social isolation, and stigma around for...

arxiv.org

It's critical for independent researchers and oversight bodies to be able to thoroughly assess and audit chatbot behavior, D&S researcher @briana-v.bsky.social says; right now AI labs are “grading their own homework.” spectrum.ieee.org/mental-healt...

Guardrails for Chatbots Aim to Protect Hearts And Minds

Design, testing, and functionality could protect vulnerable users

spectrum.ieee.org

I am at #CHI2026 representing @datasociety.bsky.social! I'm participating in the Social & Emotional Uses of AI workshop, where I'll be contributing findings from our longitudinal study on how people use LLMs for mental & emotional care. If you're around, please say hi! @chi.acm.org

🚀Introducing 𝐆𝐔𝐈𝐃𝐄-𝐋𝐋𝐌: A reporting checklist for using LLMs in behavioral & social science ✅GUIDE-LLM is a reporting checklist designed by 80+ experts to improve transparency, reproducibility & ethical accountability of LLM-based research 📄 llm-checklist.com

While many people feel relief after using chatbots for mental health support, they are designed to keep you coming back and can drive dependence. "You need to really be careful about what forms of other care you’re displacing,” says D&S researcher @briana-v.bsky.social. www.msn.com/en-us/health...

Feb 26: Join researchers @livgar.bsky.social & @briana-v.bsky.social, in conversation with Luca Belli, @mbogen.bsky.social & @marlynnweimd.bsky.social, as they explore ongoing research on mental health and chatbots, and how to navigate a profound shift in care. datasociety.net/events/menta...

Next week @livgar.bsky.social and I are hosting a discussion on our research into how people use LLMs for care/support—with voices across psychiatry, policy, engineering, and research in one room (rare & urgently needed). Come ask hard questions + share thoughts 💙 datasociety.net/events/menta...

Mental Health, Chatbots, and the Future of Care

datasociety.net

Can #AI chatbots become sources of care and intimacy? This #OpenAccess study examines #MentalHealth and spiritual chatbots, revealing how users share deeply personal information while raising questions about privacy and regulation. bit.ly/3KI5C2p @tamigraph.bsky.social @alicetiara.bsky.social

A chatbot for the soul: mental health care, privacy, and intimacy in AI-based conversational agents - Communication and Change

Artificial intelligence-based conversational agents—chatbots—are increasingly integrated into telehealth platforms, employee wellness programs, and mobile applications to address structural gaps in mental health care. While these chatbots promise accessibility, they are often deployed without sufficient impact assessment or even basic user testing. This paper presents a case study using community red-teaming exercises to evaluate a chatbot designed for wellness and spirituality. Unlike traditional red-teaming, which is conducted by engineers to assess vulnerabilities, community red-teaming treats impacted users as experts, uncovering concerns related to privacy, ethics, and functionality. Our fieldwork, conducted with undergraduate beta testers (n = 28), revealed that participants were often more comfortable sharing private information with the chatbot than with a stranger. Prior experience with commercial AI systems, such as ChatGPT, contributed to this ease. However, participants also raised concerns about misinformation, inadequate guardrails for sensitive topics, data security, and dependency. Despite these risks, users remained open to the chatbot’s potential as a spiritual wellness guide. We further examine how algorithmic impact assessments (AIAs) both capture and overlook key aspects of the spiritual chatbot user experience. These chatbots offer hyper-personalized, AI-mediated divination and wellness interactions, blurring the boundaries between astrology, mental health support, and spiritual guidance. The chameleon-like nature of these technologies challenges conventional assessment frameworks, necessitating a nuanced approach that considers specific use cases, potential user bases, and indirect impacts on broader communities. We argue that AIAs for care and wellness chatbots must account for these complexities, ensuring ethical deployment and mitigating harm.

bit.ly

In our comment to the FDA, we draw on our ongoing research to focus on what people’s actual, everyday use of chatbots for mental and emotional support means for the FDA’s approach to generative AI-enabled digital mental health medical devices. datasociety.net/announcement...

Comment to the FDA on Generative AI-Enabled Digital Mental Health Medical Devices

datasociety.net

In a new Points piece, D&S researchers @livgar.bsky.social and @briana-v.bsky.social reflect on what studying AI chatbots reveals about emotional attachment to machines, the power of illusion, and the norms and expectations that define care itself. datasociety.net/points/all-t...

Researching the use of chatbots for mental health support, D&S's @briana-v.bsky.social heard from many people who expected less stigma from chatbots than from therapists, “which is really interesting, given what we know about machine-learning bias,” she tells @undark.org. undark.org/2025/11/04/c...

Researchers Weigh the Use of AI for Mental Health

Chatbots weren't designed for mental health, but they are increasingly used for therapy. What are the risks an benefits?

undark.org

🚨NEW PAPER ALERT 🚨 @tamigraph.bsky.social @briana-v.bsky.social and I discuss wellness chatbots, their labor implications, & "private vibes" vs "privacy". We call for critical human-AI communication scholarship that takes into account context & political economy: link.springer.com/article/10.1...

A chatbot for the soul: mental health care, privacy, and intimacy in AI-based conversational agents - Communication and Change

Artificial intelligence-based conversational agents—chatbots—are increasingly integrated into telehealth platforms, employee wellness programs, and mobile applications to address structural gaps in me...

link.springer.com

“The perennial availability of AI is rapidly reshaping our expectations of care,” @briana-v.bsky.social and Ranjit Singh write in a new blog post about this piece. “Care is not mere output. It lives in the slow, difficult work of attunement, which no machine can truly provide.”

Data & Society@datasociety.bsky.social · last yr.

In @bigdatasoc.bsky.social, D&S researcher @briana-v.bsky.social & program director Ranjit Singh consider how the growing adoption of AI in mental health support impacts both users & providers, and advocate for a critical approach to its development & governance. journals.sagepub.com/doi/10.1177/...

In tandem, Ranjit and I published a short guest piece that was released today on the BD&S blog! bigdatasoc.blogspot.com/2025/10/gues...

Data & Society@datasociety.bsky.social · last yr.

In @bigdatasoc.bsky.social, D&S researcher @briana-v.bsky.social & program director Ranjit Singh consider how the growing adoption of AI in mental health support impacts both users & providers, and advocate for a critical approach to its development & governance. journals.sagepub.com/doi/10.1177/...

Research (including that from tech companies) suggests that users are putting too much trust in AI chatbots. D&S researcher @briana-v.bsky.social notes that some people take chatbot output at face value, without critically considering the text the algorithms produce. undark.org/2025/09/12/c...

Are We Offloading Critical Thinking to Chatbots?

Research, much if it by companies with deep investment in AI, suggests that chatbot interactions alter how users think.

undark.org

Btw @jeffreymoro.com, @briana-v.bsky.social + I have a forthcoming special issue on Algorithms + the occult w/ @xrw.bsky.social @drs.bsky.social AX Mina @thechristinet.bsky.social @jessrauchberg.bsky.social @aketchum22.bsky.social @serife.bsky.social @emmaquilty.bsky.social Xandro Segade + more

When D&S researchers Ranjit Singh, @livgar.bsky.social‬, @briana-v.bsky.social‬, and @emnetspeaks.bsky.social‬ set out to interview people about their use of mental health chatbots, they encountered — and were fascinated by — a particular kind of AI-enabled deception. datasociety.net/points/chatb...

Excited to present "Towards AI Accountability Infrastructure: Gaps and Opportunities in AI Audit Tooling" at #CHI2025 tomorrow(today)! 🗓 Tue, 29 Apr | 9:48–10:00 AM JST (Mon, 28 Apr | 8:48–9:00 PM ET) 📍 G401 (Pacifico North 4F) 📄 dl.acm.org/doi/10.1145/...

🎉 Participants Needed for Paid Research! Do you use chatbots (e.g. ChatGPT) for mental health? @datasociety.bsky.social is studying how people use them for support. Details: ✔️ 4-week diary study ✔️ Optional 1-hr focus group ✔️ 45-min interview 💰 Earn up tp to $400 Learn more/sign up: bit.ly/4iCLkTd

Screening Form: Chatbots for Therapy

We are conducting a research study on how people use AI chatbots for mental health support and are looking for participants! If you regularly use an AI chatbot for emotional or mental health support, ...

bit.ly

This program fundamentally changed the course of my career, and I HIGHLY HIGHLY recommend it for eligible undergraduates! Feel free to reach out to me for any tips or questions 😊

jake hofman@jakehofman.bsky.social · 2y ago

Attention NYC undergrads: Applications are open for our 12th annual Data Science Summer school at Microsoft Research NYC! Apply here by April 15th: bit.ly/3pCQENh

Just a heads-up that our X account — where we haven't posted since November of last year — appears to have been hacked, this after two of our posts were falsely flagged as a "policy violation." Good times over there! Please note that here and LinkedIn are the only platforms we're active on.

Shoutout to our all-star co-authors: Ranjit Singh, @carolanderson.bsky.social, @emnetspeaks.bsky.social, @briana-v.bsky.social, @bduckles.bsky.social, and Jacob Metcalf.

Red-Teaming in the Public Interest

This report offers a vision for red-teaming in the public interest: a process that goes beyond system-centric testing of already built systems to consider the full range of ways the public can be invo...

datasociety.net

On Thursday, Feb 20 @ 1 pm ET, Lama Ahmad, @camillefrancois.bsky.social, Tarleton Gillespie, @briana-v.bsky.social, and @borhane.bsky.social will examine red-teaming’s place in the evolving landscape of genAI evaluation and governance. RSVP and join us online! datasociety.net/events/red-t...

🚨BREAKING. From a program officer at the National Science Foundation, a list of keywords that can cause a grant to be pulled. I will be sharing screenshots of these keywords along with a decision tree. Please share widely. This is a crisis for academic freedom & science.

D&S researchers @livgar.bsky.social & Briana Vecchione are organizing a panel at @4sweb.bsky.social in Seattle called “Know Th/AI Self: Mental Health, Therapeutic Divination, and Chatbot Conspirituality” (see #177). Apply now: the CFP closes on Friday! www.4sonline.org/accepted_ope...