“It's not even close to being ready to replace anything”: the perception of AI use in youth mental health services (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Artificial intelligence (AI) technology has made significant advancements in healthcare. A key application of AI in mental health is the use of AI-powered chatbots; however, empirical evidence on their effectiveness remains limited. </sec> <sec> <title>OBJECTIVE</title> This study explored stakeholder perceptions of integrating AI in youth mental health services, focusing on its potential benefits, challenges, usefulness, and regulatory implications. </sec> <sec> <title>METHODS</title> This qualitative study utilized semi-structured in-depth interviews with 23 mobile health stakeholders, including youth users, service providers, and non-clinical staff from an integrated youth service network. We used an inductive approach and thematic analysis to identify and summarize common themes and sub-themes. </sec> <sec> <title>RESULTS</title> Participants identified AIH's potential to support education, navigation, and administrative tasks in healthcare, as well as to create safe spaces and mitigate health resource burdens. However, they expressed concerns about the lack of human elements, such as empathy and clinical judgment. Key challenges included privacy issues, unknown risks from rapid technological advancements, and insufficient crisis management for sensitive mental health cases. Participants viewed AIH's ability to mimic human behavior as a critical quality standard and emphasized the need for a robust evaluation framework combining objective metrics with subjective insights. </sec> <sec> <title>CONCLUSIONS</title> While AIH has the potential to improve healthcare access and experience, it cannot address all mental health challenges and may exacerbate existing issues. Even AIH can complement less-complex services, it cannot replace the therapeutic value of human interaction at this time. Co-design with end-users is critical for successful AI integration. Robust evaluation frameworks and an iterative approach to build a learning health system are essential to refine AIH and ensure it aligns with real-world evolving needs. </sec>

Ähnliche Arbeiten