Chancen und Herausforderungen von KI-basierten Chatbots zur Unterstützung der psychologischen Athletenbetreuung

Abstract

Elite athletes show elevated rates of psychological distress but seek help less often due to stigma, time constraints, and structural barriers. This thesis examines the extent to which AI-supported chatbots can meaningfully complement psychological care in elite sport. The aim is a context-sensitive transfer assessment using the PIET-T model. Following an overview review in sport psychology (with no robust evidence on chatbot systems), a systematic search was conducted for corresponding applications in general psychology (January–June 2025). From 46 publications, 17 full-text studies were included for the PIET-T transfer analysis. The four PIET-T dimensions are reported across these studies with varying depth. Based on the original criteria catalogue, a short form with 25 core criteria was developed. The studies were coded on three levels and combined into equally weighted scores, supplemented by a qualitative appraisal. Results: “Intervention” (≈ 75%) and “Transfer process” (≈ 80%) are well documented; “Population” reaches ≈ 65%. The largest gap is the “Environment” (≈ 48%): data protection and role allocation, as well as organizational embedding, are often insufficiently addressed. Short-term, small-to-moderate improvements in anxiety/depression scores and high initial acceptance are evident, alongside typical usage drop-offs. Opportunities: location- and time-independent availability; lowthreshold, less stigmatizing access; structured self-reflection and longitudinal monitoring; relief of scarce professional resources; good scalability and adaptability to training and travel rhythms. Challenges: clear data sovereignty and consent flows; rolebased access rights (athlete, coach, psychologist); travel robustness (offline buffering, secure synchronization); lack of long-term evidence; sport-specific content; binding leadership and team communication, as well as exit and vendor-switch concepts. Conversational AI chatbots have substantial but conditional potential. Successful transfer requires six prerequisites: fit with the target group, sport-specific modules while preserving content fidelity, role-based data protection, hybrid embedding into existing care pathways, offline robustness, and season-spanning evaluation with fixed metrics and audit cycles. The assessment relies predominantly on studies with short follow-up periods; cultural and motivational aspects are often insufficiently reported. In addition, single-coder procedures and the lack of field testing in elite-sport settings limit the conclusiveness of the findings.

Date of Award	2025
Original language	German (Austria)
Awarding Institution	Johannes Kepler University Linz
Supervisor	René Riedl (Supervisor)