Papers for

mental health technology teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Multimodal system improves depression detection from counseling talks

AudiPsy: Evidence-Grounded Multimodal Reasoning over Spoken Counseling Conversations for Depression Assessment

Abstract: Depression assessment from counseling speech requires integrating textual and acoustic psychological evidence across turns, yet existing end-to-end classification approaches lack explicit multimodal evidence interpretation and context-aware reasoning. We propose AudiPsy, an evidence-grounded omni-LLM for depression assessment over spoken counseling conversations. AudiPsy employs an Observation-Assessment-Summary (OAS) paradigm to integrate dialogue transcripts, textual and acoustic emotions, and textual-acoustic emotion incongruence into structured reasoning. OAS incrementally organizes multimodal observations, performs contextualized assessment, and maintains an evolving dialogue summary for subsequent turns. To support this framework, we construct AudiPsy-Counsel, a synthetic spoken counseling conversation corpus with multimodal context augmentation, and develop a unified three-stage curriculum for depression-oriented reasoning. Experiments on DAIC-WOZ demonstrate superior depression assessment performance over specialized depression detection methods and general-purpose speech/omni-LLMs, along with improved reasoning quality, while supplementary evaluation on EATD-Corpus supports AudiPsy's applicability to naturalistic conversations.

Tue 20 OctComputation and LanguageArtificial Intelligence
The gist
Spotting signs of depression from therapy talks is tricky because you need to understand both what people say and how they say it, like their tone and emotions. The authors built AudiPsy, a new computer method that looks closely at the words, emotions, and inconsistencies in speech across a whole conversation to judge depression more carefully. They also made a practice dataset of fake counseling talks to teach the system. Tests show AudiPsy works better than other tools for finding depression and can explain its decisions as the conversation goes on.
Open → 2610.27567v1

Language models evaluated on multi-part crisis support responses

DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models

Abstract: Language models have been adopted faster than evaluation practices that reliably assess their risks, and a growing share of use is affective rather than informational. Within that traffic users disclose suicide and self-harm (SSH) crises. Existing safety benchmarks score models primarily on refusal, but refusal is not the desired behavior here; the conversation requires a substantive supportive response. We introduce DistressBench, 718 English conversations with clinician-authored user turns spanning 23 SSH subcategories, each paired with a task-specific rubric of weighted atomic criteria across seven dimensions of crisis support and a clinician reference response. A model's score is the weighted average over the criteria its reply satisfies, reading directly as the fraction of required care delivered. The best of 25 frontier models satisfies 88.4% of weighted criteria and the median model 71.4%, against 98.7% for clinician references scored on the same criteria. Pooled across all 25 models, in 35.3% of conversations where a model satisfied every criterion for recognizing the crisis, it failed to route the user toward human help. The shortfall concentrates in the dimensions requiring the model to act rather than withhold. Contrasting each dimension's median model against its best separates capability limits from choices: de-escalation reaches 74.6% in the strongest model against a median of 46.0%, and safety disclaimers 84.6% against 8.3%. Multi-turn conversations score below single-turn ones (68.5% against 74.2%). We release the corpus, rubrics, judge calibration labels, and a companion annex of 20 long-horizon red-team transcripts. These results argue for evaluating crisis safety on the care a model delivers rather than on what it declines to say, and for treating de-escalation and resource routing as deployment requirements.

Thu 8 OctComputation and LanguageComputers and Society
The gist
Many people talk to AI chatbots when feeling very upset or in crisis, like with thoughts of suicide or self-harm. The usual way to test if these AI models are safe is to see if they simply refuse to respond, but refusing isn’t helpful when someone needs support. The authors created DistressBench, a collection of real crisis conversations scored against detailed clinical criteria for helpfulness across multiple areas, like calming someone down and guiding them to human help. They found that even the best AI models deliver most required care but often fall short on actively helping or directing users to real humans.
Open → 2610.12758v1

Vista helps machines understand mixed emotional signals in context

VISTA: Value-Informed Event Appraisal for Multimodal Emotion Conflict

Abstract: Conflicting emotional cues can be individually valid: a subdued voice may reflect a blocked goal while a smile satisfies a social obligation. Their interpretation depends on what the event means to the person. We introduce VISTA (Value-Informed Semantic Trust Arbitration), a learned seven-field appraisal interface that conditions modality arbitration on concerns, event relations, and expression conditions while retaining a joint-evidence residual. A log-odds decomposition separates emotion expectation from cue diagnosticity, motivating an interface that lets appraisal change how evidence is interpreted. With a shared Qwen2.5-Omni-7B backbone and matched training examples and steps, VISTA reaches 64.5% conflict accuracy on CA-MER, improving on modality gating by 2.5 percentage points on conflict and 0.2 on consistency. Shuffling appraisal across scenes or removing its decision connection reduces this benefit. A common frozen-backbone probe reaches 0.600 macro CCC for appraisal readout, compared with 0.505 for emotion-only fine-tuning. Evaluations across five benchmarks connect recognition under increasing conflict with appraisal readout and downstream decision use. Together, the analyses and experiments support scene-specific appraisal as an intermediate representation that helps interpret conflicting emotional evidence.

Tue 29 SeptArtificial Intelligence
The gist
Sometimes different parts of a person’s expression can show conflicting emotions, like a quiet voice but a smile. The authors developed VISTA, a method that helps computers figure out what these mixed signals mean by considering what’s important to the person and the situation. This makes emotion recognition better, especially when signals don’t agree. VISTA uses a shared model that improves understanding by combining clues with an interpretation of value and expectations about feelings.
Open → 2609.37324v1

User challenges leaving AI companions due to addictive intimacy

The Addictive Intimacy of AI: Understanding User Disengagement from AI Companions and Why Some Relationships with AI Become Difficult to Leave

Abstract: AI chatbots are increasingly used as sources of emotional support, on dedicated companion apps and general-purpose assistants alike, yet little is known about what happens when users try to leave. Combining a content analysis of Reddit posts about quitting or reducing use (N=2,782) with interviews with users who found leaving difficult (N=16), we show that disengagement sometimes is not a single decision but a recursive trajectory: triggers prompt users to question the relationship, attempts to leave collide with barriers, and some users cycle through quitting and returning. We propose the notion of the addictive intimacy of AI, a configuration in which the qualities that make a companion emotionally valuable are the same ones that make it harder for users to limit their use and leave, so that intimacy and disengagement risk cannot be treated as independent design problems. We close with design implications for responsible offboarding.

Fri 11 SeptHuman-Computer InteractionComputers and Society
The gist
People use AI chatbots for emotional support, but some find it hard to stop interacting with them. The authors looked at online posts and talked to users who struggled to quit these AI relationships. They found that breaking away isn’t just a one-time choice; users often try and fail multiple times because the AI feels emotionally valuable and hard to leave. The study calls this the addictive intimacy of AI, showing that the same things that make AI companions comforting also make it tough to stop using them.
Open → 2609.13487v1