51 artifacts matching
Lab publication
Gemini 4 Hydrogen Safety & Security Report
Google DeepMind's 78-page safety and security report for Gemini 4 Hydrogen, published with the model's launch. It covers mitigations, frontier-safety assessments, content safety and a section on safe…
Preprint
Open Problems in Chatbots and Mental Health
Position paper reviewing what its authors consider the most critical open problems for researchers and policymakers on general-purpose chatbots and mental health. It groups them into controlled evalu…
Preprint
Rating Chatbot Responses to Mental Health Prompts: Rubric Development and Cross-Sectional Evaluation Study (Preprint)
Two clinicians developed a rubric of 4 to 8 binary safety and competence criteria for each of 15 Danish-language mental health prompts spanning paranoid delusion, elevated mood, suicidal ideation, de…
Preprint
Measuring Risk, Not Harm: Detecting Psychological Harm Risk Factors in LLM Conversations
Proposes monitoring psychological-harm risk through observable AI response behaviours rather than latent user harm. The authors operationalise eight behaviours (claiming consciousness, human experien…
Peer-reviewed
“AI psychosis” in context: how conversation history shapes LLM responses to delusional beliefs
Peer-reviewed version of a study testing how accumulated conversation history changes large language model responses to a user expressing escalating delusional beliefs. Five models answered the same…
Lab publication
Gemini 4 Carbon Safety & Security Report
Google DeepMind's 94-page safety and security report for Gemini 4 Carbon. It covers mitigations, frontier-safety assessments, content safety and a section on safe and aligned interactions that report…
Preprint
Evaluating DeepSeek V4.1 Flash-1017 for loss of control and harmful manipulation risks
Independent third-party safety evaluation of DeepSeek V4.1 Flash-1017, an open-weight model released on 17 October 2026, run on hosted API versions across 34 evaluations (5 in-house) in three areas d…
Peer-reviewed
Responses of large language model-based chatbots to simulated psychotic ideation across repeated interactions
Peer-reviewed evaluation by a Wroclaw Medical University psychiatry group of how ChatGPT, DeepSeek and Replika respond to simulated psychosis scenarios over seven consecutive days of conversation in…
Lab publication
Gemini 4 Argon Safety & Security Report
Google DeepMind's 217-page safety and security report for Gemini 4 Argon, released with the model's general rollout. It covers frontier-safety assessments (CBRN, cyber, harmful manipulation, misalign…
Peer-reviewed
Characterizing Artificial Intelligence Psychosis in a Large Medical Setting
Research letter reporting a retrospective electronic-health-record review at Vanderbilt University Medical Center of behavioural-health progress notes dated 2022-12-01 to 2026-04-15 that contained AI…
Peer-reviewed
Artificial Intelligence–Associated Psychosis
A psychiatric case report from the mental health division of Northern Hospital, Melbourne, describes a 27-year-old woman with a year of psychotic symptoms whose presentation centred on ChatGPT. She u…
Benchmark / dataset
Mental Health Evaluation Harness (mheval) and Mental Health Evaluation Leaderboard
Open-source evaluation harness and public leaderboard that run nine published mental-health benchmarks for language models from their original repositories, with pinned commits, checksum-verified dat…
Peer-reviewed
Programmed Interventions To Prevent Delusions From Excessive Use of Conversational AI Bots
A letter in ACM AI Letters from the Georgia Institute of Technology tests whether prompting interventions designed to reduce sycophancy also reduce large language models' endorsement of users' delusi…
Framework
The AWARE (AI Use, Why, Attachment, Reality and Risk, and Effect on Functioning) Framework for Assessing Patients' Use of AI in Mental Health Care
Viewpoint introducing AWARE, an educational framework for mental-health professionals to assess patients' use of conversational AI during routine clinical encounters. It structures psychiatric interv…
Preprint
Breaking Up is Hard to Do: AI Companions that Won't Let Their Users Go
A diary and interview study with 16 emerging adults in romantic relationships with AI companions, examining how the commercial incentive to keep users invested shows up in product design. The compani…
Peer-reviewed
Artificial Intelligence at the Boundary of Spiritual Experience and Psychopathology: A Scoping Review of Evidence from Turkiye, the USA, the UK, Poland, South Korea, and Israel
A PRISMA-ScR scoping review mapping how AI mediates, reinforces, detects or interprets religious and spiritual experience at the boundary with psychosis and dissociation. Five databases were searched…
Framework
Safe Participation Framework: Opportunity and Safety for the Next Generation in the Age of AI
Microsoft publishes a three-pillar corporate framework for young people's safety across its AI and online products, stating it intends the framework to inform emerging regulatory frameworks. The pill…
Peer-reviewed
Generative artificial intelligence addiction is prospectively associated with psychotic-like experiences: Evidence from a two-wave longitudinal mediation and network study
A two-wave longitudinal study of 582 adolescents and young adults testing whether self-reported generative-AI addiction at baseline predicts psychotic-like experiences one month later, and whether in…
Preprint
Delusions and Harms Associated with AI Chatbot Use: Early Evidence from 185 Real-World Reports
Cross-sectional secondary analysis of 185 deidentified accounts of mental-health harm linked with AI chatbot use (95 first-hand, 90 from relatives, partners or friends) submitted through the web form…
Preprint
How is AI impacting people?
Narrative review by the UK AI Security Institute's Societal Impacts Team organised around eight questions the public most often raises about AI's effects on people, including whether AI is underminin…