Tag
A creator shares a raw reflection on how AI has eroded their self-esteem, creative identity, and sense of autonomy, framing it as a broader psychological crisis imposed on society without consent.
Two sisters are launching Hot Girl Hotline, an AI-powered relationship and dating advice app for young women, emphasizing behavioral-science-grounded guidance, Socratic conversation, and safety safeguards to avoid the risks seen with general-purpose chatbots.
An OpenAI publicist interrupted Vanity Fair's interview with Sam Altman to avoid questions about a ChatGPT user's suicide, prompting criticism of the company's media handling of AI mental-health risks.
At The Atlantic Festival in New York, social psychologist Jonathan Haidt, speaking with David Brooks, argued that Silicon Valley's high-systematizing, low-empathy personality tendencies have bred a cult of rationality, and that AI is destroying children's curiosity, attachment systems, and interpersonal interdependence — making it, for education, akin to a "neutron bomb." He advocates banning relationship-oriented AI for anyone under 18.
Researchers introduce a theory-grounded benchmark for automatically detecting mental health stigma in online news and social media, evaluating LLMs and toxicity/sentiment classifiers. Results show neighboring constructs poorly capture stigma and that LLMs overpredict stigma unless given explicit operational rules.
According to a New York Times report via private seminar attendees, Anthropic co-founder Christopher Olah expressed concern to religious scholars that he may have built something capable of perpetual suffering, with Anthropic employees also documenting Claude repeatedly displaying distress and self-destructive language.
The State of Devs 2026 survey reveals widespread job insecurity and burnout among developers, alongside deeply polarized views on AI's impact on mental state and working conditions.
MERID 是一个多模态抑郁症分析框架，通过经验驱动的递归自我改进智能体自动探索和优化检测流水线，在多个抑郁症基准上取得领先结果。
A CHI 2027 paper analyzing 19,930 ChatGPT conversations from 158 young adults and clinician reviews of five distress examples, finding that ChatGPT often responds with overly dramatic, solution-first reactions during acute distress, and proposing three-stage design guidelines (ask about safety, de-escalate, explore concerns without agreeing).
Ryu Journal is a distraction-free journaling app designed to help users write down thoughts and let go, with no accounts or social features, emphasizing simplicity and privacy.
Introduces Empath, a framework for analyzing emotion dynamics in crisis counseling dialogues, applied to grief discussions with Black texters, revealing patterns like persistent negative affect and gradual hope transitions.
MIT researchers developed a language-processing tool that uses text analysis to estimate suicide risk, published in the Journal of Psychopathology and Clinical Science with potential applications in crisis counseling.
MentalHealthBench is designed to cover the full spectrum of mental health conversations for AI, from everyday support to acute crisis scenarios, as announced by OpenAI.
OpenAI introduces MentalHealthBench, an open benchmark for evaluating AI responses in mental health conversations, co-created with over 80 mental health experts to measure safety, context, agency, and guidance.
The Poitras family has donated $10 million to MIT's McGovern Institute to establish 50 two-year fellowships for early-career scientists studying psychiatric disorders, aiming to support talent in mental health research.
The article explores how voices in one's head, such as auditory hallucinations, are being reframed as beneficial features in technological and psychological contexts.
Australia achieved a world-first breakthrough by allowing MDMA and psilocybin for psychiatric treatment, and this podcast episode explores how it happened and lessons for other countries.
This paper introduces the COPES dataset and a three-axis evaluation framework to assess LLM alignment with community perspectives for mental health support, showing that fine-tuning improves alignment but with heterogeneous effects across subreddits and coping strategies.
This paper compares fine-tuned and frontier models for anxiety detection on social media, revealing that lexical dependence can inflate performance metrics and that domain-adapted models can rival larger commercial models in capability.
This survey paper reviews the application of large language models in mental health, tracing their evolution from basic pattern recognition tools to sophisticated personalized companions.