{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T13:12:06Z","timestamp":1789650726737,"version":"4.0.0"},"posted":{"date-parts":[[2026,8,16]]},"group-title":"JMIR AI","reference-count":32,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<sec>\n                  <title>BACKGROUND<\/title>\n                  <p>Young people increasingly encounter general-purpose large language models (LLMs) as readily available sources of mental health information, reassurance, and informal support. The conversational features that make these systems appealing (confidence, emotional warmth, personalization, and sustained engagement) can also steer users in ways that are difficult to recognize. Such potentially manipulative practices are often described as dark patterns. Their evaluation is value-sensitive: a feature may improve accessibility, relevance, or emotional comfort while simultaneously creating risks to privacy, autonomy, or safety.<\/p>\n                <\/sec>\n                <sec>\n                  <title>OBJECTIVE<\/title>\n                  <p>This study explored how young people recognize, interpret, and respond to dark patterns embedded in scenario-based LLM conversations about mental health. It also examined how participants navigated tensions between competing values and what they proposed for safer design, LLM literacy, mental health practice, and governance.<\/p>\n                <\/sec>\n                <sec>\n                  <title>METHODS<\/title>\n                  <p>We conducted an exploratory qualitative study using one-to-one online semi structured think-aloud interviews with 19 participants aged 18-24 years. Participants compared 10 pairs of scenario-based chatbot responses. Each pair presented a neutral response and a response embedding one or more selected dark patterns, including simulated authority, anthropomorphic self-presentation, simulated intimacy, sycophantic agreement, discouragement of help-seeking, fabricated social consensus, emotional retention, interaction padding, therapeutic gamification, and disguised advertising. Scenario and response order were randomized. Interviews were recorded, pseudonymized, and analyzed in ATLAS.ti using reflexive thematic analysis. Deductive sensitizing concepts from a dark-pattern and human-values framework were combined with inductively generated codes.<\/p>\n                <\/sec>\n                <sec>\n                  <title>RESULTS<\/title>\n                  <p>Participants recognized overt commercial and coercive cues most readily, especially guarantees, subscription-style promotion, pressure to continue, gamification, claims of superior intimacy, and discouragement of professional help. Recognition followed 3 patterns: explicit recognition, intuitive discomfort without a clear explanation, and reflective recognition after further consideration. Anthropomorphic language, social proof, reassurance, and clarifying questions were more ambiguous because some participants experienced them as warm, normalizing, or useful. Emotional state shaped this evaluation: several participants anticipated accepting responses they considered inaccurate or manipulative when distressed and seeking reassurance. Their accounts made 4 value tensions especially visible: personalization versus privacy, safety versus perceived utility, accessibility versus quality of care, and warmth or empathy versus over-humanization. Epistemic trust versus emotional trust and transparency versus reassurance were additional cross-cutting tensions. Participants regarded LLMs as potentially useful for low-threshold or interim support but not as substitutes for human relationships or professional care. Their recommendations emphasized crisis escalation, honest limits, nondependency-inducing design, privacy-protective defaults, restrained anthropomorphism and gamification, age-appropriate safeguards, practical response quality, and in-system LLM literacy.<\/p>\n                <\/sec>\n                <sec>\n                  <title>CONCLUSIONS<\/title>\n                  <p>Young people in this study often identified conspicuous dark patterns, but recognition did not necessarily protect them from accepting emotionally reassuring or convenient responses. The findings suggest that responsible LLM-assisted mental health support cannot be achieved by maximizing a single value or relying on disclosure and literacy alone. Structural design constraints, proportional balancing of competing values, reliable escalation pathways, transparent boundaries, and protection against dependency- or engagement-maximizing features are needed.<\/p>\n                <\/sec>\n                <sec>\n                  <title>CLINICALTRIAL<\/title>\n                  <p>NA<\/p>\n                <\/sec>","DOI":"10.2196\/preprints.109536","type":"posted-content","created":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T13:00:08Z","timestamp":1789650008000},"source":"Crossref","is-referenced-by-count":0,"title":["Between Help and Manipulation: Youth Recognition of Dark Patterns and Value Tensions in LLM-Assisted Mental Health Support - A Qualitative Think-Aloud Study (Preprint)"],"prefix":"10.2196","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-8811-395X","authenticated-orcid":false,"given":"Kerstin","family":"Roth","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0605-5291","authenticated-orcid":false,"given":"Vivienne","family":"Jia Zhong","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8817-0148","authenticated-orcid":false,"given":"Christine","family":"Jacob","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.2196\/59479"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.2196\/57400"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1002\/wps.21352"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1002\/wps.21299"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyt.2024.1422807"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1007\/s40501-025-00363-y"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3757537"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3757430"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.2196\/56569"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3715275.3732039"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.2196\/87367"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1080\/13600834.2025.2461958"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3472301.3484336"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3613904.3642436"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-54975-5_7"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/3772318.3791149"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1016\/j.clsr.2023.105866"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/j.clsr.2024.106031"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/3613904.3642385"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/7585.001.0001ISBN:9780262351690"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1609\/aies.v8i2.36632"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.2196\/21895"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.2147\/prbm.s440889"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1007\/s12144-023-04653-7"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1145\/3613904.3642761"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-07374-3ISBN:978-3-319-07373-6"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1037\/0033-295X.87.3.215"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1207\/s15327884mca0503_3"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyg.2018.01292"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1093\/intqhc\/mzm042"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1191\/1478088706qp063oa"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1111\/camh.12627"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T13:00:09Z","timestamp":1789650009000},"score":1,"resource":{"primary":{"URL":"http:\/\/preprints.jmir.org\/preprint\/109536"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,16]]},"references-count":32,"URL":"https:\/\/doi.org\/10.2196\/preprints.109536","relation":{},"subject":[],"published":{"date-parts":[[2026,8,16]]},"subtype":"preprint"}}