The Impact of Generative Conversational Artificial Intelligence on the Lesbian, Gay, Bisexual, Transgender, and Queer Community: Scoping Review

Abstract

Background Despite recent significant strides toward acceptance, inclusion, and equality, members of the lesbian, gay, bisexual, transgender, and queer (LGBTQ) community still face alarming mental health disparities, being almost 3 times more likely to experience depression, anxiety, and suicidal thoughts than their heterosexual counterparts. These unique psychological challenges are due to discrimination, stigmatization, and identity-related struggles and can potentially benefit from generative conversational artificial intelligence (AI). As the latest advancement in AI, conversational agents and chatbots can imitate human conversation and support mental health, fostering diversity and inclusivity, combating stigma, and countering discrimination. In contrast, if not properly designed, they can perpetuate exclusion and inequities. Objective This study aims to examine the impact of generative conversational AI on the LGBTQ community. Methods This study was designed as a scoping review. Four electronic scholarly databases (Scopus, Embase, Web of Science, and MEDLINE via PubMed) and gray literature (Google Scholar) were consulted from inception without any language restrictions. Original studies focusing on the LGBTQ community or counselors working with this community exposed to chatbots and AI-enhanced internet-based platforms and exploring the feasibility, acceptance, or effectiveness of AI-enhanced tools were deemed eligible. The findings were reported in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews). Results Seven applications (HIVST-Chatbot, TelePrEP Navigator, Amanda Selfie, Crisis Contact Simulator, REALbot, Tough Talks, and Queer AI) were included and reviewed. The chatbots and internet-based assistants identified served various purposes: (1) to identify LGBTQ individuals at risk of suicide or contracting HIV or other sexually transmitted infections, (2) to provide resources to LGBTQ youth from underserved areas, (3) facilitate HIV status disclosure to sex partners, and (4) develop training role-play personas encompassing the diverse experiences and intersecting identities of LGBTQ youth to educate counselors. The use of generative conversational AI for the LGBTQ community is still in its early stages. Initial studies have found that deploying chatbots is feasible and well received, with high ratings for usability and user satisfaction. However, there is room for improvement in terms of the content provided and making conversations more engaging and interactive. Many of these studies used small sample sizes and short-term interventions measuring limited outcomes. Conclusions Generative conversational AI holds promise, but further development and formal evaluation are needed, including studies with larger samples, longer interventions, and randomized trials to compare different content, delivery methods, and dissemination platforms. In addition, a focus on engagement with behavioral objectives is essential to advance this field. The findings have broad practical implications, highlighting that AI’s impact spans various aspects of people’s lives. Assessing AI’s impact on diverse communities and adopting diversity-aware and intersectional approaches can help shape AI’s positive impact on society as a whole.

Defining acceptable data collection and reuse standards for queer artificial intelligence research in mental health: protocol for the online PARQAIR-MH Delphi study

“… 64 65 We will need to develop anticipated use-cases and scenarios that might predictably arise in the application of data-driven technology in mental healthcare (or healthcare more generally) which are sparse—a recent exception being the application of conversational AI to provide mental health support for the LGBTQI+ community. 66 …”

“…64 65 We will need to develop anticipated use-cases and scenarios that might predictably arise in the application of data-driven technology in mental healthcare (or healthcare more generally) which are sparse-a recent exception being the application of conversational AI to provide mental health support for the LGBTQI+ community. 66 For both targeted reviews, two reviewers from the executive group will independently screen titles/ abstracts (for the research literature) or 'executive summaries' or landing pages (for web-based resources) to identify eligible artefacts for full review and inclusion. Reviewers will first attempt to resolve disagreements through discussion; if discussion fails to resolve disagreements, a third reviewer will break ties.…”

More questions than answers: Ethical considerations at the intersection of psychology and generative artificial intelligence.

“…These forms of genAI-mediated treatment have been explored for addressing pervasive psychological concerns such as depression and anxiety (Abd-Alrazaq et al, 2019; Fiske et al, 2019) and used to implement several interventions from evidence-based treatments (e.g., acceptance and commitment therapy; Martinengo et al, 2022; cognitive behavioral therapy; Ly et al, 2017; dialectical behavior therapy; Dimeff et al, 2021). GenAI-mediated psychotherapeutic interventions have been praised for decreasing clinician burden and increasing access to services among largely underserved and vulnerable populations (Bragazzi et al, 2023; Sedlakova & Trachsel, 2023), as these services may be available when practitioners are not (Sweeney et al, 2021). GenAI may also be used to support clinical or diagnostic decision-making (Giordano et al, 2021, Magrabi et al, 2019) and review ongoing symptom monitoring to improve treatment outcomes (Davenport & Kalakota, 2019).…”