Evaluating the Experience of LGBTQ+ People Using Large Language Model Based Chatbots for Mental Health Support
Open accessThe DOI matches its Crossref record; 40 of the 40 references it cites were found in a registry or an index.
Abstract
LGBTQ+ individuals are increasingly turning to chatbots powered by large language models (LLMs) to meet their mental health needs. However, little research has explored whether these chatbots can adequately and safely provide tailored support for this demographic. We interviewed 18 LGBTQ+ and 13 non-LGBTQ+ participants about their experiences with LLM-based chatbots for mental health needs. LGBTQ+ participants relied on these chatbots for mental health support, likely due to an absence of support in real life. Notably, while LLMs offer prompt support, they frequently fall short in grasping the nuances of LGBTQ-specific challenges. Although fine-tuning LLMs to address LGBTQ+ needs can be a step in the right direction, it isn’t the panacea. The deeper issue is entrenched in societal discrimination. Consequently, we call on future researchers and designers to look beyond mere technical refinements and advocate for holistic strategies that confront and counteract the societal biases burdening the LGBTQ+ community.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
1 retracted
Darcy, A., Beaudette, A., Chiauzzi, E., Daniels, J., Goodwin, K., Mariano, T. Y., Wicks, P., & Robinson, A. (2022) · Expert Review of Medical Devices · doi:10.1080/17434440.2022.2075726 · Cited by 40
RetractedChatGPT for Good? On Opportunities and Challenges of Large Language Models for Education
Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., et al. (2023) · doi:10.35542/osf.io/5er8f · Cited by 7,552
ChatGPT for good? On opportunities and challenges of large language models for education
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., et al. (2023) · Learning and Individual Differences · doi:10.1016/j.lindif.2023.102274 · Cited by 7,326 · cited 2×
“The advanced linguistic capabilities of LLM-based chatbots offer users more context-aware and responsive interactions, distinguishing them from the earlier pre-LLM chatbots [56].”
BackgroundOn the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445922 · Cited by 6,988 · cited 4×
“LGBTQ+ users, who are often underrepresented in the training datasets, can encounter unintentional reinforcement of damaging narratives with regard to their identities [9, 34].”
BackgroundInfluentialELIZA—a computer program for the study of natural language communication between man and machine
Weizenbaum, J. (1966) · Communications of the ACM · doi:10.1145/365153.365168 · Cited by 4,452 · cited 1×
“Rule-based chatbots operate on predefined rules, linking user inputs to specific responses [110].”
BackgroundMinority Stress and Mental Health in Gay Men
Meyer, I. H. (1995) · Journal of Health and Social Behavior · doi:10.2307/2137286 · Cited by 4,264 · cited 1×
“Minority stress theory suggests that structural stigma against LGBTQ+ people, interpersonal discrimination, and internalized stigma all exacerbate the mental health challenges of this population, resulting in feelings of alienation and distress [22, 47, 76].”
BackgroundFitzpatrick, K. K., Darcy, A., & Vierhile, M. (2017) · JMIR Mental Health · doi:10.2196/mental.7785 · Cited by 2,730 · cited 3×
“Although LLM-based chatbots seem to offer a valuable and inclusive mental health resource for the LGBTQ+ community, potentially bridging gaps in traditional therapy accessibility [37, 69], their deployment raises substantial concerns.”
BackgroundA method of analysing interview transcripts in qualitative research
Burnard, P. (1991) · Nurse Education Today · doi:10.1016/0260-6917(91)90009-y · Cited by 2,510
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé, H., & Crawford, K. (2018) · arXiv · doi:10.48550/arxiv.1803.09010 · Cited by 2,474
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., III, H. D., & Crawford, K. (2021) · Communications of the ACM · doi:10.1145/3458723 · Cited by 2,469 · cited 1×
“LGBTQ+ participants’experienceinusingthechatbotsshowsthatthegenericpurposessofLLMstrainedonlargecorpusmightnotbeinclusive—howthedataiscollected ,annotated,andused,aswellaswhois involvedinthecurationanddesigningprocessescanhavesignif-icantimplicationsforLGBTQ+users[40].”
Advantages and limitations of Internet-based interventions for common mental disorders
Andersson, G., & Titov, N. (2014) · World Psychiatry · doi:10.1002/wps.20083 · Cited by 1,160
Chatbots and Conversational Agents in Mental Health: A Review of the Psychiatric Landscape
Vaidyam, A. N., Wisniewski, H., Halamka, J. D., Kashavan, M. S., & Torous, J. B. (2019) · The Canadian Journal of Psychiatry · doi:10.1177/0706743719828977 · Cited by 1,048 · cited 1×
“…drawbacks, such as limited linguistic or contextual understanding, often led to unnatural or irrelevant conversations, reducing users’ willingness to engage with these chatbots, making interactions less convincing and supportive, and potentially limiting therapeutic benefits [11, 60, 82, 105].”
BackgroundArtificial intelligence for mental health and mental illnesses: an overview
Graham, S., Depp, C., Lee, E. E., Nebeker, C., Tu, X., Kim, H.-C., & Jeste, D. V. (2019) · Current Psychiatry Reports · doi:10.1007/s11920-019-1094-0 · Cited by 970
Fiske, A., Henningsen, P., & Buyx, A. (2019) · Journal of Medical Internet Research · doi:10.2196/13216 · Cited by 857 · cited 2×
“Designers should also seek to ensure the balance between user autonomy within the chatbot interface and facilitating timely access to safety resources [35, 42].”
BackgroundHollis, C., Falconer, C. J., Martin, J. L., Whittington, C., Stockton, S., Glazebrook, C., & Davies, E. B. (2017) · Journal of Child Psychology and Psychiatry · doi:10.1111/jcpp.12663 · Cited by 846 · cited 1×
“By providing self-guided, affordable, accessible, and private mental health care, they address key barriers to traditional therapy, including long waiting times, extended treatment duration, and traveling costs [5, 50, 72].”
BackgroundFulmer, R., Joerin, A., Gentile, B., Lakerink, L., & Rauws, M. (2018) · doi:10.2196/preprints.9782 · Cited by 715
RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models
Gehman, S., Gururangan, S., Sap, M., Choi, Y., & Smith, N. A. (2020) · Findings of the Association for Computational Linguistics: EMNLP 2020 · doi:10.18653/v1/2020.findings-emnlp.301 · Cited by 693
Beede, E., Baylor, E., Hersch, F., Iurchenko, A., Wilcox, L., Ruamviboonsuk, P., & Vardoulakis, L. M. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376718 · Cited by 614
An overview of the features of chatbots in mental health: A scoping review
Abd-alrazaq, A. A., Alajlani, M., Alalwan, A. A., Bewick, B. M., Gardner, P., & Househ, M. (2019) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2019.103978 · Cited by 605 · cited 1×
“Research indicats that mental health chatbots have had positive impacts in reducing symptoms of depression and anxiety, and enhancing therapeutic alliance, acceptability, and like-ability, particularly during the COVID-19 pandemic [1, 2, 45, 81, 97].”
BackgroundUser Experiences of Social Support From Companion Chatbots in Everyday Contexts: Thematic Analysis
Ta, V., Griffith, C., Boatfield, C., Wang, X., Civitello, M., Bader, H., DeCero, E., & Loggarakis, A. (2020) · Journal of Medical Internet Research · doi:10.2196/16235 · Cited by 561 · cited 1×
“Especially noteworthy is the safe environment for intimate conversations these chatbots provided to LGBTQ+ participants, mirroring previous use with that of the pre-LLM chatbots [39, 99].”
BackgroundAbd-Alrazaq, A. A., Rababeh, A., Alajlani, M., Bewick, B. M., & Househ, M. (2020) · Journal of Medical Internet Research · doi:10.2196/16021 · Cited by 503 · cited 1×
“Research indicats that mental health chatbots have had positive impacts in reducing symptoms of depression and anxiety, and enhancing therapeutic alliance, acceptability, and like-ability, particularly during the COVID-19 pandemic [1, 2, 45, 81, 97].”
BackgroundArtificially intelligent chatbots in digital mental health interventions: a review
Boucher, E. M., Harake, N. R., Ward, H. E., Stoeckl, S. E., Vargas, J., Minkel, J., Parks, A. C., & Zilca, R. (2021) · Expert Review of Medical Devices · doi:10.1080/17434440.2021.2013200 · Cited by 440
The Personalization of Conversational Agents in Health Care: Systematic Review
Kocaballi, A. B., Berkovsky, S., Quiroz, J. C., Laranjo, L., Tong, H. L., Rezazadegan, D., Briatore, A., & Coiera, E. (2019) · Journal of Medical Internet Research · doi:10.2196/15360 · Cited by 332
Contending Stories: Narrative in Social Movements
Polletta, F. (1998) · Qualitative Sociology · doi:10.1023/a:1023332410633 · Cited by 309 · cited 1×
“Older Wikipedia pages have been revised to provide more cohesive narratives over time, impacting the data that shapes LLMs [83].”
Challenges in building intelligent open-domain dialog systems
Huang, M., Zhu, X., & Gao, J. (2020) · ACM Transactions on Information Systems · doi:10.1145/3383123 · Cited by 288
Enacted Stigma, Mental Health, and Protective Factors Among Transgender Youth in Canada
Veale, J. F., Peter, T., Travers, R., & Saewyc, E. M. (2017) · Transgender Health · doi:10.1089/trgh.2017.0031 · Cited by 259 · cited 2×
“LGBTQ+ Individuals The LGBTQ+ community experiences greater mental health challenges such as higher levels of depressive symptoms, engaging in more non-suicidal-self-injury, and having more suicidal thoughts and behaviors compared with heterosexual, cisgender peers [4, 53, 89, 100, 101, 107].”
BackgroundA Survey on In-context Learning
Dong, Q., Li, L., Dai, D., Zheng, C., Ma, J., Li, R., Xia, H., Xu, J., Wu, Z., Liu, T., et al. (2023) · arXiv · doi:10.48550/arxiv.2301.00234 · Cited by 248
For You, or For"You"?: Everyday LGBTQ+ Encounters with TikTok
Simpson, E., & Semaan, B. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3432951 · Cited by 230
Maintaining reality: Relational agents for antipsychotic medication adherence
Bickmore, T. W., Puskar, K., Schlenk, E. A., Pfeifer, L. M., & Sereika, S. M. (2010) · Interacting with Computers · doi:10.1016/j.intcom.2010.02.001 · Cited by 213
Pizzi, G., Vannucci, V., Mazzoli, V., & Donvito, R. (2023) · Psychology & Marketing · doi:10.1002/mar.21813 · Cited by 194
Street-Level Algorithms: A Theory at the Gaps Between Policy and Decisions
Alkhatib, A., & Bernstein, M. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300760 · Cited by 188 · cited 2×
“A case in point is YouTube, where content from trans individuals discussing their gender and sexuality has faced demonetization [3].”
BackgroundHe, Y., Yang, L., Zhu, X., Wu, B., Zhang, S., Qian, C., & Tian, T. (2022) · Journal of Medical Internet Research · doi:10.2196/40719 · Cited by 178
Jacobs, M., He, J., F. Pradier, M., Lam, B., Ahn, A. C., McCoy, T. H., Perlis, R. H., Doshi-Velez, F., & Gajos, K. Z. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445385 · Cited by 177
Self-citationDesigning Trans Technology: Defining Challenges and Envisioning Community-Centered Solutions
Haimson, O. L., Gorrell, D., Starks, D. L., & Weinger, Z. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376669 · Cited by 136 · cited 1×
“However, online technologies can sometimes fall short of meeting the needs of the LGBTQ+ community, as they do not center LGBTQ+ people in the design process [43].”
BackgroundA Mathematical Investigation of Hallucination and Creativity in GPT Models
Lee, M. (2023) · Mathematics · doi:10.3390/math11102320 · Cited by 105 · cited 1×
“Furthermore, LLMs’ propensity for generating hallucinated responses can mislead or confuse users [61, 64].”
BackgroundCooper, R. M., & Blumenfeld, W. J. (2012) · Journal of LGBT Youth · doi:10.1080/19361653.2011.649616 · Cited by 99
WinoQueer: A community-in-the-loop benchmark for anti-LGBTQ+ bias in large language models
Felkner, V., Chang, H.-C. H., Jang, E., & May, J. (2023) · Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2023.acl-long.507 · Cited by 74
Black Lives Matter in Wikipedia: Collective Memory and Collaboration around Online Social Movements
Twyman, M., Keegan, B. C., & Shaw, A. (2017) · Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing · doi:10.1145/2998181.2998232 · Cited by 51 · cited 2×
“Even when the filtered versions are used, they may inadvertently offend and silence the voices of marginalized communities such as LGBTQ+ due to inherent limitations in the filtering algorithms [104].”
BackgroundMa, Z., Mei, Y., & Su, Z. (2023) · arXiv · doi:10.48550/arxiv.2307.15810 · Cited by 40
Self-citationWilcox, L., Shelby, R., Veeraraghavan, R., Haimson, O. L., Erickson, G. C., Turken, M., & Gulotta, R. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581040 · Cited by 34
Cited by
Private Yet Social: How LLM Chatbots Support and Challenge Eating Disorder Recovery
Choi, R., Kim, T., Park, S., Kim, J. G., & Lee, S.-J. (2025) · Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3713485 · Cited by 19
Therapy as an NLP Task: Comparing LLMs and Human Peers Behaviors in CBT Sessions
Iftikhar, Z., Ransom, S., Xiao, A., Nugent, N., & Huang, J. (2024) · arXiv · doi:10.48550/arxiv.2409.02244 · Cited by 6
Skierś, K., Peng, D., Kundu, A., Person, T., Takase, K., Nagoshi, T., Nakayama, S., Yuichiro, Y., Miyazaki, T., Minamizawa, K., et al. (2025) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2025.3616754 · Cited by 5
Butorac, I., McNaney, R., Thomas, J., Flores-Herrera, A., Seguin, J. P., Northam, J. C., & Carter, A. (2026) · ACM Transactions on Computing for Healthcare · doi:10.1145/3805040 · Cited by 3
Artificial intelligence and immersive technologies for LGBTQ+ youth: a narrative review
Shekarchi, R., Eleuteri, S., Saladino, V., Calaresi, D., Giordano, F., & Prunas, A. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.2002054
Metadata
- Authors
- Ma, Z., Mei, Y., Long, Y., Su, Z., & Gajos, K. Z.
- Venue
- Proceedings of the CHI Conference on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2024-05-11
- Pages
- 1-15
- OpenAlex
- W4391835981
- Cited by
- 68
- References
- 40
- FWCI
- 28.03
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.