Exploring Gender Bias in AI for Personalized Medicine: Focus Group Study With Trans Community Members.

Background: This paper explores the perception and application of artificial intelligence (AI) for personalized medicine within the trans community, an often-overlooked demographic in the broader scope of precision medicine. Despite growing advancements in AI-driven health care solutions, little research has been dedicated to understanding how these technologies can be tailored to meet the unique health care needs of trans individuals. Addressing this gap is crucial for ensuring that precision medicine is genuinely inclusive and effective for all populations. Objective: This study aimed to identify the specific challenges, obstacles, and potential solutions associated with the deployment of AI technologies in the development of personalized medicine for trans people. This research emphasizes a trans-inclusive and multidisciplinary perspective, highlighting the importance of cultural competence and community engagement in the design and implementation of AI-driven health care solutions. Methods: A communicative methodology was applied in this study, prioritizing the active involvement of end-users and stakeholders through egalitarian dialogue that recognizes and values cultural intelligence. The methodological design included iterative consultations with trans community representatives to cocreate the research workflow and adapt data collection instruments accordingly. This participatory approach ensured that the perspectives and lived experiences of trans individuals were integral to the research process. Data collection was conducted through 3 focus groups with 16 trans adults, aimed at discussing the challenges, risks, and transformative potential of AI in precision medicine. Results: Analysis of the focus group discussions revealed several critical barriers impacting the integration of AI in personalized medicine for trans people, including concerns around data privacy, biases in algorithmic decision-making, and the lack of tailored health care data reflective of trans experiences. Participants expressed apprehensions about potential misdiagnoses or inappropriate treatments due to cisnormative data models. However, they also identified opportunities for AI to enhance health care outcomes, advocating for community-led data collection initiatives and improved algorithmic transparency. Proposed solutions included enhancing datasets with trans-specific health markers, incorporating community voices in AI development processes, and prioritizing ethical frameworks that respect gender diversity. Conclusions: This study underscores the necessity for a trans-inclusive approach to precision medicine, facilitated by AI technologies that are sensitive to the health care needs and lived realities of trans people. By addressing the identified challenges and adopting community-driven solutions, AI has the potential to bridge existing health care gaps and improve the quality of life for trans individuals. This research contributes to the growing discourse on equitable health care innovation, calling for more inclusive AI design practices that extend the benefits of precision medicine to marginalized communities.